© Roger M Tagg 2008, 2010
Welcome to FROLIO – a new attempt to merge philosophy and the "semantic web". This website is in process of continuous development.
In many fields of human knowledge, a simplifying assumption is often made that things are formed into a single, simple hierarchy.
Literally, a hierarchy is an organization structure of priests. Junior priests are subject to the command of higher ranking priests. These are in turn subject to bishops, who report to archbishops and possible a single supreme pope. Such a relationship is almost universally "one to many" - many subordinates are subject to only one superior. This reflects the view that "one cannot serve two masters".
This "subject to the command of" relationship is mirrored in armed forces and many business or government organizations. However over time, it has become realized that strict hierarchies may have disadvantages, e.g. lack of flexibility in fast changing circumstances. A popular alternative is "matrix management". Typically, each person may have two bosses; one who manages the group of people with a common skill, the other who manages projects for a particular class of customers, as in the example below.
You can see a good real-world example at the website of Eurocontrol, the EU's air traffic management authority.
Of course, this has disadvantages too (for a discussion read this paper). Things can become messy. Many humans prefer to be in a simple hierarchy, just as they prefer to stick with the customs and religion of their forefathers. Those in positions of command may also prefer to feel that they have total control over the individuals that report to them. Possibly this is all bound up historically with what's called the Enlightenment (see also this, and my own views).
As part of the Enlightenment, Carl Linnaeus, a Swedish scientist, developed a hierarchical classification of all living things. This is rather different from a reports to hierarchy, since it is concerned with classes or types of living thing, rather than individual persons or management positions. The one to many relationship here is is a specialization of. Valuable and long-lasting though Linnaeus's classification system is, it is not without problems. Linnaeus based his hierarchy on structural similarity. Later, Darwin's theory of evolution came along, and later still, DNA and genome analysis; both these caused modifications, and exposed shortcomings of a pure hierarchy.
As an example, consider the existence of Australian mouse-like marsupial rodents; they may be functionally and structurally similar to Eurasian mice in many ways; but would appear far removed in the hierarchy. There is also the problem of hybridisation, where types that have not diverged too far can re-pool their genes and create a new form, e.g. the Adelaide Rosella (see near the bottom of Wikipedia's article on Crimson Rosella) or Leyland Cypress. After all, many living species need both a male and a female in order to reproduce, so that the fundamental process is "two to many" rather than "one to many".
Things get very much messier when we try to draw up hierarchical classifications of artificial things, like mechanical or electrical parts. We can organize by several different principles, e.g. shape, composition, purpose, containment, fitting etc.
Another messy area lies in the words and phrases used in a spoken or written language. Ways of expressing thoughts, observations, questions etc are extremely varied. The concept of a Thesaurus has been developed to record relationships between different words and phrases. A thesaurus can show broader and narrower terms, synonyms etc, but there are many other possible relationships such as common roots, common derivation, underlying meaning etc. A Dictionary can only order words alphabetically (or in Chinese, by number of strokes).
A lot of effort is currently being put in by information specialists to formalise knowledge and meaning, primarily for use with electronic computing. Because of the need to check their "models" with the people who actually do things in the real world, diagrams are used as an aid to discussion. There have been very many types of diagrams used, with name such as flowcharts, function hierarchies, entity-relationship diagrams, state transition diagrams and object models.
The subject of "Informatics" (alias "IT"), like religion, is very much subject to factions, each wanting to push their own orthodoxy. One recent IT fashion has been Object Orientation (OO for short). In this, real world things are modelled as if they are all separate small computers, each of which has its own internal memory and operations, but externally can only pass messages to each other. For this to be efficient, operations have to be shared between similar objects; so hierarchies of classes of objects (like Linnaeus's) have become the paramount view. But as a consequence, other relationships and concerns are relegated to a lower priority and may not get considered at all, with bad consequences for the computer system and its users.
There are also problems with these simple "one to many" is-a relationships. One is that they may not be one to many. In many universities, a senior student can work as a tutor, so would inherit some properties from the class of "teachers" and some from "students". Such a student is-a teacher and is-a student. This is known as multiple inheritance. Another problem is that is-a does not always mean quite the same thing. We can say that a certain individual, Fred, is-a student (meaning is an instance of) or that "third year student" is-a student (meaning is a specialization of).
Of course there may be other types of relationship as well. For example a trunk is-part-of an elephant, and a baby elephant is-offspring-of a father and mother elephant.
The conclusion is that simple hierarchies can only capture one relationship at a time, and even then may not accurately capture reality. While in some cases there may be a clear winner in which relationship should be the primary one, in many cases there is no clear winner. However if we try to show all the relevant relationships in one diagram, with the annotations needed to explain their precise meaning, we would end up with an impossible spaghetti-like tangle. What is needed is the ability to switch easily between the different views, using some sort of “Change View” button.
These link you, the reader, to more details about FROLIO. The core of this website is a set of pages about relationships - links to these are on the first table below. The index in the bottom row points to a set of "diatribes" or opinionated essays, about things the author feels particularly strongly about!
Some of these links may be under construction – or re-construction.
This version updated on 29th January 2010
If you have constructive suggestions or comments, please contact the author rogertag@tpg.com.au .