An introduction to clear semantic modelling
On this page, I introduce my notion of clear semantic modelling in the context of philosophy. In subsequent pages, I share and then analyse my own mostly grounded clear semantic model of conscious reality, and then use it to analyse the semantic model of a major philosophical work.
Philosophy is simply clear thinking about reality, about our place in it, and about how we ought to behave in it. One way to think clearly about reality is to construct clear semantic models. Semantic models, as the term implies, are models of meaning. They make use of syntactical elements - words or other symbols - with each of which is associated some stipulated or understood meaning, in order to express a key concept via a coherent set of component concepts and the relationships between them.
A clear semantic model is:
- Semantically clear:
- Piecewise distinct: Each component concept, and the meanings of the words or other symbols that it uses, do not overlap one another without good cause, where a potential good cause is that one is a useful abstraction of the other.
- Piecewise precise: Each component concept is defined clearly and with precision.
- Unequivocal: Each word or other symbol used is defined to have either a single meaning, or, if it is defined to have multiple meanings, then the different meanings in the different contexts in which the word or other symbol is used are made clear.
- Essentially denotationally consistent: The stipulated meaning of each word or symbol does not contradict the essential meaning of its common or technically contextual usage, though it might, for clarity in context, be specified more precisely than its essential common or technical meaning.
- Conceptually clear:
- Essential: The model expresses a key concept.
- Non-extraneous: The model uses no more component concepts than are necessary to express the key concept.
- Complete: The model uses sufficient component concepts to express the key concept.
- Piecewise necessitated: Each component concept is necessary to support the key concept.
- Precise: Each component concept has an appropriate degree of abstraction to the model's explanatory purpose, and, overall, the model as a relationship of component concepts expresses the key concept at an appropriate degree of abstraction.
- Relationally integrated: Each component concept is in some appropriate relationship to at least one other, and all component concepts are related to one another via the network.
- Consistent: No concept or relationship contradicts any other, and no set of concepts or relationships taken together entail a contradiction.
- Justified: A good reason can be given for the inclusion and scope of each component concept or relationship to the extent that it is not speculative.
- Parsimonious: To the extent that the model is theoretical, its theories comprise the minimum conceptual substance necessary to explain those facts which they are intended to explain.
- Optimal: The model could not express the key concept better than it does at its level of precision using alternative component concepts and/or the relationships between them.
- Representationally accurate: If the semantic model is representative, then its representation is accurate of that which it represents, otherwise this criterion does not apply.
At a meta level, this notion of clear semantic models is itself (intended to be) a clear semantic model. To the extent that it can be shown to fail to meet its own criteria for clarity, I am prepared to revise it.
Further characteristics of semantic models
Semantic models vary along an axis from entirely representative - of a reality other than themselves - to entirely self-referential, as in axiomatic mathematical systems.
🔗 An aside to define "axis" and "dimension"
An axis is a construct which entails some variable varying directionally between "farther" and "nearer", or varying between "more" and "less", or varying between "greater" and "lesser". An example of an axis is a simple geometrical line, along which the variable of "distance" varies from "nearer" to "farther" (from some fixed point) along the line, in one direction (positive) or the other (negative). The axis of semantic models just introduced is one in which the variable of "representativity" varies between more representational and less representational, or, equivalently, between less self-referential and more self-referential.
Though this term is not used on this page, I introduce it because it will be used in subsequent pages: a dimension is an axis whose ontological nature is not conceptual but is the basis upon which tangible energy is differentiated: the energy is differentiated along the axis (typically, this is in the plural form of "dimensions", and the energy is differentiated along multiple axes simultaneously, as in our three-dimensional nominally "physical" reality).
Semantic models exist on an axis from entirely grounded to entirely theoretical. Entirely grounded semantic models are those based solely in grounded facts and propose no contingent facts. Entirely theoretical semantic models propose solely contingent facts.
Theoretical semantic models also exist on an axis from entirely testable to entirely untestable or speculative. A test in this sense is an observation or experiment which tends to confirm or falsify the accuracy of the model or the truth of some ungrounded fact in which it is based.
🔗 An aside to define grounded and contingent facts, so as to elucidate the distinction between grounded versus theoretical models
Grounded facts are those which are effectively infallible. They are of three types: experientially-grounded, semantically-grounded, and deductively-grounded.
Experientially-grounded facts start with the fact of my being conscious, which is indubitable to me, and then extend into facts about the existence of the types of experiences that I have - affective, somatic, sensory-perceptual, and cognitive - the existence of which, too, are indubitable to me, and then into the specific experiences that I have: though I might be mistaken about that which, for example, a specific sensory-perceptual experience represents, I cannot be mistaken that I am experiencing that representation. For example, if I experience a lion in front of me, I might be wrong that the experience represents a lion in front of me (it might actually represent a cleverly-designed robot with the form of a lion), but I cannot be wrong that I am experiencing the representation of a lion - whatever it is that that representation turns out to represent.
Semantically-grounded facts are those based in meaning and logic: they are facts of the type that one meaningful expression is necessarily identical to, or necessarily entails, another. They are facts grounded, then, in necessary logical transformations. Semantically-grounded facts, then, assume the infallibility of logic.
Through the application of semantically-grounded facts (logical transformations which are effectively infallible) to experientially-grounded facts (the transformed essentially infallible facts), new grounded facts (the effectively infallible outcomes of the transformation) can be derived: these are deductively-grounded facts or simply deduced facts.
Contingent facts are those which are effectively fallible: they are ungrounded and we hold them to be true for reasons which are themselves ungrounded facts. Through the application of semantically-grounded facts to contingent facts, we can reliably produce inferred facts, which themselves are ungrounded.
The next page presents a mostly grounded semantic model of conscious reality.
