FROLIO – Formalizable Relationship-Oriented Language-Insensitive Ontology - Home Page

© Roger M Tagg 2007-2010

Welcome to FROLIO – a new attempt to merge philosophy and the "semantic web". This website is under continuing development.

Rationale - what is the point of all this?

Contents of this Page: Why Formalizable?    Why Language-Insensitive?    Why Relationship-Oriented?    Why Ontology?    Influences    Principles    A First Glimpse   Links    About the Author

Introduction

I’ve been suspecting, over a lifetime of 50 years as an adult, that the languages we use to communicate with our fellow humans may be beautiful and culturally distinctive, but they are beginning to show signs of strain as the complexity of our lives and our thoughts increases. Trying to think clearly using our languages has become tough. Vested interests and various hustlers have been hijacking our vocabulary in an attempt to influence or even control our minds. I’m thinking of Advertisers, Salesmen, Management, Union Activists, Politicians, Fundamentalists and anyone whose interests depend on the continuance of traditional institutions. In short, I’m fed up of being bullshitted.

OK, so maybe now I've reached retirement, I haven’t got that long to put up with this bullshit, but I aim to die fighting it. I don’t claim a lot of philosophical or linguistic background. But I have had a 47-year career in information science of one sort or another, and I have seen a number of attempts to formalize information to enable computers to do things with it.

Why Formalizable?

Most readers of this web page will know that computers can – for better or worse – do a lot with information these days. The World Wide Web links millions of people with shareable knowledge and facilities. However, searching for what is relevant, and matching what is related, is still relatively hit-and-miss. Google may represent the state of the art at present, but this amounts to little more than matching on words and strings of characters. Doing business over the web is fine, but the seller’s words don’t always match the buyer’s words, even if they are talking about the same thing. And this applies even between people who are all using the same language. So if we can construct a common understanding of what we are talking about, rather than trying to rely on words in languages all the time, computers might have more chance to help us. They may not be up to par with a bright human assistant at the moment, but I wouldn’t bet against them catching us up in many ways in the next 50 years.

Note that I say "formalizable" and not "formalized". I don't think I am anywhere near creating a universal language that resolves all doubt. Maybe if I could live another 50 years, though?

Why Language-Insensitive?

With much talk about Globalization, the “Global Village” and Global Warming, it’s clear to me that we can’t look at things only from the viewpoint of our own language or country. It’s especially hard for me – I’m now into my third citizenship and have lived in another country with a different language and a different religion. Even an English speaker like myself has to face the fact that there may now be more web pages in Chinese than in English. Even with good human and computerized translation, we may still fail to communicate the right message.

About 20 years ago I had an occasional job chairing public courses on IT which were offered by a company owned by the late great Robert Maxwell. Most of the attendees were from non-English speaking backgrounds. I recently came across a doodle I had drawn at the bottom of my notepad supplied by the company. I’ve attempted to reproduce it in the figure below.

Strings of symbols (Latin, Greek, Cyrillic, Arabic, Devanagari, Korean, Japanese, Chinese or whatever) that are recognizable as words or phrases, carry meaning in various languages. However the same string may represent two different ideas, for example in English “bananas” could mean a type of bendy yellow fruit - or “crazy”. So I added an extra level called “lexical term”, which meant the use of a string to represent one single concept - as one would find in a dictionary. Sometimes the different terms have the same spelling but different pronunciation, e.g. mate (= friend) and mate (= Paraguay tea); although in English this is often caused by borrowing from other languages.

The right hand side - which is where FROLIO concentrates - is concerned with a hopefully common understanding of what we are talking about, independent of the words used. To make sense though, all the possible "concepts" have to be put in some sort of structure. This structure includes the common recognition of "relationships" between the "concepts". We also suppose that each language would need a definition of each "concept" – so we can never be 100% independent of language. But potentially, reliable speakers of each language that can understand the concepts can get together to cross-check their definitions.

Why Relationship-Oriented?

To be precise, FROLIO is based on inter-concept relationships (not on inter-string relationships like puns, anagrams that might be relevant to a crossword addict). I am hoping to avoid too much dependence on the syntax structure of any one language, e.g. nouns, verbs and adjectives in English. Although verbs do sometimes represent “actions”, taking this as the rule might bias the structure too much to action and not enough to thought. In some languages, common verbs like “is” and “has” are not directly represented (e.g. Arabic). I have chosen to make relationships a key structuring concept as I feel that this is what we as humans experience, whether it is human-to-human, human-to-things or things-to-things.

I also favour an approach to relationships which is not limited to two things, or even two kinds of things. My relationships may have many “roles” – this is explained on the Roles in Rels page.

Around 30 years ago I was in a discussion on the lines of whether such a thing as the “real world” really exists. A colleague was arguing that his particular model of information was geared to representing the real world. I remember saying – without any great confidence – that I thought that the real world was just “the consensus of observed relationships”. Maybe what I am doing now is experimenting to see if that view makes any sense.

Why Ontology?

I am using this term in the sense currently used by the information specialists who are building the so-called “semantic web”. An ontology is simply a structure of meaning, either within a particular domain of knowledge, or at the level of talking about knowledge generally. The semantic web idea comes from Tim Berners-Lee, the originator of the web itself. The point is that, in spite of the open nature of the web, we cannot be sure that if two strings of characters are the same, they mean the same thing. We are still missing a common, shared structure of meaning and knowledge. Many present day developments in ontology use a computer encoding system known as OWL (Ontology Web Language), and in my own research work I have been using this as far as I can. Whether it is totally adequate I am yet to discover.

Hasn’t someone else done it all before?

Some philosophers in the past tried to build complete "systems" in an attempt to cover everything that we can understand about the world. Three notable attempts are Leibniz's "monads", Kant's categories and Hegel's dialectics. These have had their followers, Marx having developed his ideas from Hegel. However the old philosopher nearest to my own approach is Dogen, the Japanese Zen master from as long ago as 1250 AD, who said "Things do not have a meaning in themselves, but only in relation to other things". Richard Osborne, a recent interpreter of arcane philosophies for the man in the street, commented "For Westerners obsessed with the classification of material things, this is a truly radical idea".

Historically, not all philosophies have had a clear separation between language and concepts. Wittgenstein said words to the effect that “when we refer to objects in the world we are operating within a language game, and our concept of ‘the world’ and how we divide up the world is already conditioned by the structure of our language” (according to John Searle of UC Berkeley, in Magee “The Great Philosophers” BBC Press, 1987). However my view is that we’ve got lots of different languages in the world and we need to understand each other better, so we can’t just leave it there. As Neurath (of the Vienna circle) said, "words divide, pictures unite" - so maybe he would not criticize my use of clip art.

In my view, the real hero is Peter Mark Roget, who published the first edition of his Thesaurus in 1852. Roget was for a time Secretary to the British Royal Society, an august learned institution crossing many disciplines. At the time that I drew the doodle mentioned earlier, I was tempted to christen my project “Roger’s Thesaurus”, but I wanted to separate it a bit further from the particular language of English. Also, in those days, I knew what a Thesaurus was but hadn’t heard of such a thing as an Ontology. Since I semi-retired in 2007, I have had time to follow it up.

The interesting structure in Roget is not so much the alphabetic index of words and phrases, but the structural arrangement by concept. Roget’s basic scheme is shown below (from Penguin Roget’s Thesaurus).

• Class 1: Abstract Relations

• Class 2: Space

• Class 3: Matter

• Class 4: Intellect: the exercise of the mind

• Division one: Formation of ideas

• Division two: Communication of ideas

• Class 5: Volition: the exercise of the will

• Division one: Individual volition

• Division two: Social volition

• Class 6: Emotion, religion and morality

There is a further division into 39 sub headings before the basic list of around a thousand individual concept entries, each with a barrage of words and phrases, split into nouns, verbs, adjectives etc. The words and phrases are a mixture of synonyms, related terms, aspects, examples and so on.

Roget did not attempt to be any of “Formalizable”, "Relationship-Oriented" or "Language Insensitive". Of course his work was done well before the era of today's computers. He also admitted that his categorization could not succeed in being perfect, because some concepts can be regarded as subordinate to more than one category. Such a shortcoming is inherent to printed books and early computer files which are forced to be essentially sequential. But with modern databases and hypertext this limitation no longer applies. The only problem is that many ontologists and informaticians seem over-influenced by the fashion for "object-oriented" approaches to information, which do not emphasize relationships other than "is-a" (by which we mean that one class of things (e.g. elephants) is more specialized than a more general class of things (e.g. animals). Therefore, I believe, some analysts are too tied to single hierarchical classifications, and have not yet got to grips with the many different semantics that appear in classification and aggregation relationships. To find out more, you could read articles and papers about Mereology, which is the science of "part-of" relationships (definitely not the same as "is-a").

The mainstream equivalents of my project lie in such specialist areas as what are called the Dublin Core and Upper Ontology. I expect to overlap with some of this work, although my aim is not to be constrained by their current assumptions.

Another stalwart of this field is Barbara Ann Kipfer, who as well as editing a version of Roget, has also published "The Order of Things" Random House, revised 2001.

Some guiding principles – or prejudices?

I do not want this work to be seen as claiming some revealed truth about the real world. I am only too aware that it is the theory of one individual. But it is the model of one who listens rather than talks, who is impatient and cynical, but who wants to make his own mind up and not be forced to take on anyone else’s preconceptions. Some would say - an archetypal Grumpy Old Man.

Some concepts I try to avoid over-using are: Fact, Truth and Common Sense. This is not because I oppose them, but because they are words that have been sorely misused and hi-jacked. I prefer an approach based around the idea of “consensus”. I regard assertions, rather than being true or false, as being more or less well generally accepted. I tend to be suspicious of anyone who says “the facts are” or “the truth is”, since they probably mean “my interpretation is”. I think Pontius Pilate (who asked Jesus Christ “what is truth?”) may have been more of a philosopher than he is given credit for. I prefer to talk about assertions being “universally accepted” – like the earth being an oblate spheroid - or “true by definition” like 2 plus 2 equalling four. I hardly need to remind the reader that there was a time when it was “universally accepted” that the world was flat, and that the earth was the centre of the universe..

What about Common Sense? It’s a bit like Fact and Truth – people who talk about it are often implying that anyone who doesn’t agree with their viewpoint is lacking common sense. But from a pure linguistic viewpoint, it should mean the same as “consensus” – in other words, what most people (but possibly in a limited part of the whole human population) agree with.

Without apology, I am not tying this work to any particular religion or "biblical" authority. I don't regard myself as an atheist - I believe that the word "God" represents a pervasive spirit of Good. But having lived for over two years in an Islamic country - and having worked or travelled in India, Japan, China and Korea - I cannot believe that any subset of humanity has a monopoly on Good - or on revelation of it. I don’t personally want to be tied to traditional - or mythological - patterns of thought and authority. I am also (like the Chinese philosopher Wang Shou-Jen) suspicious of religions that emphasize detachment and don't front up to the material issues that face us.

I feel that we cannot disregard advances in human knowledge and understanding because they contradict traditional beliefs. According to Alan Finkel’s editorial in Cosmos magazine (issue 12, Dec 2006/Jan 2007), there is a difference between "belief" (a thought package that we accept and do not regard as subject to change) and "theory" (something that can be tested, and if and when it fails, someone will propose something better). FROLIO falls clearly into the "theory" category.

Finally, I see my goal as somewhat similar to that of John Locke, which was "to be employed as an under-labourer in clearing the ground a little, and removing some of the rubbish that lies in the way of knowledge".

A coarse categorization of relationships

The table below shows examples of how some of the different types of relationship that we can observe could be placed into categories. However I don't want to force too strict a hierarchy. Many relationships have characteristics of more than one "parent" category, just as the natural world contains hybrids.

Links

Explanations of major components in FROLIO

Some of these links may be under construction – or re-construction.

About the Author

Roger Tagg was born in Oundle, England in 1941. He attended schools locally, in Cambridge and at Mill Hill, north London. He graduated in Mathematics at Cambridge University in 1962. He then worked in Operations Research and Systems Analysis in the UK National Coal Board (later British Coal). He worked for over 20 years as a consultant in computer systems and information management, initially for a company and latterly as a freelance. More recently he has been a lecturer in computing and information systems, 7 years in New Zealand and since 2000 at the University of South Australia, from which he retired in December 2009.

He has worked abroad in Europe, the Middle East and the Far East, and has travelled widely in most continents. His grandfather and an uncle were ministers of religion, and his father - a Maths teacher - was also a lay preacher. Roger has been married to Susanne (Sue) since 1969.