The Insight-Centric Knowledge Model — ICKMOD

The “Insight-Centric Knowledge Model” (ICKMOD) is my personal take on several key aspects of a model for representation of meaning. I expect disagreement from others — even on the basics. But it’s often easier to start by negotiating differences with an existing proposal.

See also, What are the characteristics of a good model for practical knowledge?

We usually try to communicate “insights”

When we decide to say something or write a note, it’s usually because the thought is an insight of some sort — not a list of facts, but a new understanding that may have resulted from putting facts and observations together.

In knowledge work, we create value by communicating these insights – by expressing the relationships among things in reality that are important and relevant to a stated problem or opportunity. No, not natural-language statements but the meaning behind those natural-language statements and other forms of communication, including graphics, audio, and video. Words and pictures do not, by themselves, create value.

We start with meaning. Nothing constructive happens until we turn information into meaning. That’s a habit we have forgotten, in a sense, as literacy became a goal rather than one of many possible means to the goal of representing and communicating meaning … and universal publishing became a reality.

In this blog, I will use the term Insight (with an initial capital letter) to indicate that I am using the term in the sense explained in this blog post.

Insights are the basic unit of exchange of meaning, the currency of human creation of value. Cause and effect, influence, membership, and conditions are some of the relationships among the things that are relevant to us. We express Insights in natural language, but we need to move beyond natural language — or more accurately, to roll back our over-dependence on natural language in order to bring forward the meaning behind Insights.

An Insight-centric model supports the way we work

We cannot apply good practices and technology effectively to a resource if that resource does not have a well-understood model. Applying technology to the superficial properties of an un-modeled resource gives you a very small, marginal — and unpredictable — return on investment. Compare the necessity to adopt standardization of parts before moving to assembly lines.

There is an area of technology labeled “idea management.” There are “argumentation-support systems.” There are “decision-support systems.” And there are a growing number of applications — including “semantic search” — that are designed to make retrieval and integration of information more effective. These technologies have lessons to offer, but none of them truly addresses what we do in knowledge work in a direct, coherent, and comprehensive way.

So, I present ICKMOD — the “Insight-Centric Knowledge Model” — which provides the following:

  1. An appropriate deconstruction of what happens in the creation, sharing, evolution, and management of human knowledge in common work situations.
  2. Support for creating an explicit memory of the development of meaning by individuals and groups and the transfer of meaning among members of groups. This model enables individuals to describe the specific, fine-grained activities of members of the organization as they relate to creation of value — again, in well-defined, manageable formats — as structured “insights.” This transformation partially supplants the haphazard, costly, and unproductive use of unstructured communications of many kinds, including text-based documents, email, and unguided collaboration. But it is also a complement to those inevitable processes.
  3. A coherent, scalable model for explicit representations of the realities of the organization — including the market entities, resources, constraints (standards and regulations), and other factors — in formats similar to those already used in the fields of knowledge representation (KR), process modeling, argumentation support, and project management — so that those representations can be managed, retrieved, compared, evaluated, and applied in predictable ways.
  4. An object/relationship model that lends itself to graphic treatments, in effect off-loading memory of complex realities to computers, leaving many high-level aspects of creating, integrating, remembering, and managing knowledge to the parts of our brains that naturally support such activities.
  5. Infrastructure for identifying concepts, facts, and insights that are the same, similar, or related.
  6. A target for incremental formalization. Add anything at any time. Connect and formalize as time and inspiration permit.
  7. An overarching vision that enables integration of many existing practices and technologies that affect the development, management, and application of “knowledge.”

ICKMOD is an “informal abstract reference model”

My short-term goal is to produce — in conjunction with those who have similar interests — an “informal abstract reference model” for the domain of representing practical knowledge. This model focuses on how people interpret and refine the meaning of human communications — not on enabling computers to interpret natural language. Not that there’s anything wrong with that. (See also, Reference model in Wikipedia.)

ICKMOD does not suggest or require a formalism for knowledge representation. In fact, it is important to guard against premature formalization. Nevertheless, I want people to start laying the groundwork for representing their own “domains” of knowledge, however large or small, as soon as possible, using a few basic principles.

Such a model should also draw from and ultimately, I hope, enrich the domain of Knowledge Representation, in which several formalisms for “knowledge representation” have been proposed, and from other disciplines, technologies, and practices. (See my graphic about disciplines that influence ICKMOD.)

With luck, those working on ideas and technologies in these disciplines will be able to point to such a model and say, “That feature or capability, defined as precisely as possible, is what I am talking about.” Or even, “What I am talking about is currently beyond the scope of that model, but it deserves a place in that model.”

A hierarchy of “knowledge objects” in ICKMOD

In this blog, I will use the following hierarchy of objects, starting from “simple” objects of meaning to complex complex objects. (In all following posts, I will try to stick to the convention of capitalizing the first letter of these terms when they refer to the special sense in which they are used here.)

  • Concepts are the smallest building blocks. A Concept is our mental image (thought) of a specific thing or class of things, as reflected in the classic “triangle of meaning.” See also, What is a “Concept”?
  • Concepts are combined via syntax into Facts. A Fact is a simple assertion (or proposition) about reality that may be meaningful to people in a wide range of contexts. Facts include Observations — simple statements about what we see, hear, or feel as we go about the business of living — and other kinds of simple assertions whose meaning can be expressed with some degree of rigor, making them amenable to parsing for validity by computer applications.
  • An Insight is, first and foremost, an idea that seems new, useful or meaningful in our communications with others — or even in our own notes to ourselves. Insights are why we communicate.A set of Facts and the relationships among them form an Insight. The “default” form of an Insight includes a node that names (and explains) the Insight — in clear natural language — plus all the Fact nodes that support that Insight.
  • Arguments, Theses, and Domains [of knowledge]. Just as Facts support Insights, Insights can, in turn, support Arguments. Clearly, any particular Fact may support many Insights, and any particular Insight may participate in larger Arguments, and so forth.

This is my own terminology and it is arbitrary, but it can be useful as a starting point.

These objects and relationships have unique identities. Any set of objects and the relationships among them are addressable, enabling human evaluators to point at them unambiguously. And it is important to remember that each of these elements can have multiple “parents.”

I will devote additional blog posts to Concepts, Facts, and Insights.

© Copyright 2017 Philip C. Murray


This entry was posted in ICKMOD, representation of practical knowledge. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *