2018-04-17 Comments on the slides and a strange proposal on context
On the strange thing/proposal first
My interest in context comes from a linguistic stance. The objective of this Ontology Summit is to understand the role of upper ontologies in providing or using context. The two verbs, namely provide and use are an indication of what is happening. You either try to find a good name and place for something existing in your interpretation in an existing ontology (therefore provide a description), or you try to verify if an item in an ontology does exist or not according to your interpretation of reality. This pair of verbs is a good match of D. Crystal's DEED paradigm, an extended view of verbal utterances that explain why you must never forget about the context of anything said, from one word to a chapter. What is tangible (apart from the fact that you do not need verbal communication to make inferences about your own physical context) however is the string of words that make up a clause or a proposition either spoken or written that serve as a description. Now any such description must be preceded by some thinking and verbalization, which in turn must be followed by interpretation of the foregoing, involving another run of brain work to make sense of what is communicated. Now you can rarely communicate just one concept (in a word form) per se, especially not so when it comes to making statements about what exists or real. You are likely to perform a number of mental operations at the end of which you come up with something selected as appropriate or fit for the purpose. The topmost concept of most upper ontologies is object or thing (entity, etc.) Why? Because it is a word with the broadest sense, the most generic meaning, hence with the power of a wildcard character to fill in place in a statement that you cannot readily identify and relate to something that you want to name or talk about. It is also a unit of one, the result of beginning to count units in a discrete fashion either in space or in time. So it is not a name, a physical identifier, but an abstract term, the result of abstraction, a mental operation. And who performs that mental operation? Every one of us. Therefore the context of the (concept of) object is the person providing and/or using it... Having said that, it must be clear, that each ontologist is in action as separate contexts for their work of art and their mental operations are the contexts that need to be synchronized. It is not about making logical inferences or obeying the rules of logic. since emotion also has a play. The essential point is making comparisons. Comparisons are made all the time and the net result catalyses communication, including so far neglected mental operations. Since the next two semantic primitives are properties (attributes) and relations, it must be noted that they do not follow the options available for the classification of objects. For one thing properties do not make sense on their own, they must be grounded in objects. This is a problem for synchronizing contexts with respect to properties as properties also come from abstractions and they are based on experience/verification/falsification with respect to which ontologists are widely divided. But what can create a turmoil is that relations are not limited to the number and forms that you are familiar with from conventional upper ontologies, logic, algebra and psychology. No, the picture is more complex than that and this is why you have clauses with a verb as a predicate that is needed for any clause to make sense. I guess I should stop here for the time being as probably many questions may be provoked. Before i take my turn again and answer them as much as I can, I am going to submit my comments to the slides that I have seen.
1. Track B: Upper Ontologies for Specifying Context Slide 4 (Words Unionized) No word on its own has any meaning. Full stop.
Slide 5 There are about twenty associations that one can have at the sight of a word (meaning excluded) those are what people mean when they say they know a word.
Slide 6 An example of using a word in a sentence does not mean that the meaning of that word is defined (not even in a dictionary).
Slide 7 Concepts are a slippery slope. Equating a word with another word and claiming that they are thereby related and calling that relation a part of/a type of/ an example of, etc. is not sufficient for defining context (in a different context either), but it is just an illustration of what is called "sense.
Slide 9 There is an indication of misunderstanding here. Human languages are context sensitive, programming languages are context-free. This means that you cannot get rid of many-to-many and many-to-one and one-to many relations present in context sensitive languages. Whereas your ideal of one-tom-one relation of context-free languages is only provided at machine code level, in the code table. Nevertheless it would be very useful to build an NL dictionary by making use of two features of natural languages, namely recursion and tautology. But that would still not make such an inventory an ontology, because current dictionaries contain words that do not have referents, belong to poetry, fiction, religion, imagination, lies, etc that have a different domain of existence. That raises the issue of how to include the concept of time as existence as a default is present tense... albeit with arbitrary edges pointing in the future and the past.
Slide 10 The breakdown of Thing is not convincing for me. I have an alternative decomposition to be explained elsewhere or later.
Slide 11. it is the outcome of a labeling exercise. There is no such a thing as an independent thing. But suppose there is (ding and sich), then it is an object. When deemed to be relative, it s an object with a property (see slide). When such a thing is put in context (mediating thing), it is just rewritten with the same property. In concreto: a person, an employed person (an employee), a person in employment. In fact, in containment relation with a symbolic space (another object) /snip for the time being.
2. Track+B2 Outcomes
Slide 35 Everything as Context - The issue is one-to-many relations. If object (entity in focus) is one, the number of properties of that object (still in context)may be too high to enumerate. Conversely, the number of objects that display a particular property is again very high. The inverse relation of intension and extension in terms of quantitative description is limited for practical reason. Even the cause and effect chain is arbitrarily cut off at beginning and ending points to enable reactive action (identify a point of measurement which is always here and now. Think of the changing concept of dying subject to different measurement aspects or perspectives. Or when is a fetus considered to be a person, etc.) Context is clearly another object that an object in focus is contained by. There is only one genuine relation between objects, namely containment that you cannot observe by identifying yourself as being both inside and outside at the same time. Thus awareness, consciousness is the context of your thoughts and perceptions which call for some basic notions of orientations to be counted as a living person.
From an ontological verification issue or for the vital purpose of synchronizing with context you have the toolkit of a NL to define the particulars of your complex relations with context. Anyone speaking that NL can understand enough to see what is going on and to decide how to react. Mind you, neither people nor animals require verbalizations before they grasp the essence or "message" of the ongoing context to react physically. Survival knowledge is coded in the motor skills, no organism has time to mediate/reflect upon the input and compare it with its storehouse of various taxonomies sorted on containment relations.