This short post addresses two questions that an SFL scholar might naturally raise in response to the consolidated statement of relational ontology. The purpose here is not to revise the ontology, but to clarify how core Hallidayan insights are preserved and sharpened once the relevant ontological cuts are made explicit.
1. If potential is probabilistically weighted and updated with each instantiation, does probability not belong to meaning itself?
Halliday’s claim that linguistic potential is probabilistically weighted is both empirically grounded and theoretically powerful. It captures the fact that systems of meaning exhibit patterned tendencies, and that these tendencies shift as language is used.
The crucial clarification is ontological.
Probabilistic weighting belongs to the description of semiotic potential, not to the production of meaning. Probabilities model the distribution of past instantiations; they do not generate or drive new ones.
Each act of meaning leaves a residue that contributes to second-order patterning. Over time, this alters the statistical texture by which the system is described. In this sense, potential can be said to be “updated” with each instantiation.
What does not follow is that probability exerts causal force over meaning. Instantiation remains a perspectival cut from potential to event. Probability describes the grain of that cut after the fact; it does not supply the principle of actualisation.
Read this way, Halliday’s probabilistic system is entirely compatible with a relational ontology: probability belongs to modelling, not to meaning.
2. If language construes context as well as realising it, does this not collapse context into semantics?
Halliday’s claim that language construes context has often been misunderstood as a form of semantic idealism. In fact, it is best read as a claim about experience rather than determination.
Context exists in two analytically distinct but related ways.
First, context exists as phenomenon: situations as they are experienced, recognised, and made meaningful. In this sense, context is necessarily construed. There is no unconstrued situation, just as there is no unconstrued meaning.
Second, context exists as organised potential: historically sedimented social organisation, institutional structures, and patterned expectations. In this sense, context is not construed in the act. It conditions the act by delimiting what is recognisable, appropriate, or likely.
Language therefore both:
realises contextual potential (field, tenor, mode), and
construes contextual experience in first-order acts of meaning.
What it does not do is generate context as a determining system or collapse contextual structure into semantics. Conditioning and construal remain distinct.
3. Why These Clarifications Matter
These two questions mark the principal points at which probabilistic and contextual accounts are tempted to overreach.
By enforcing the distinctions between:
first- and second-order phenomena,
description and production,
conditioning and determination,
the relational ontology preserves Halliday’s insights while removing the ambiguity that allows reductionist readings to arise.
The result is not a departure from SFL, but a clarification of its ontological commitments: meaning remains enacted, situated, and answerable, while probability and context retain their proper explanatory roles.
Nothing essential is lost. Much confusion is avoided.
No comments:
Post a Comment