… predicting weather is rather like predicting text: one can make certain predictions about what people are going to say or write with a certain probability of being right — a probability that is significantly greater than chance, but not great enough to be easily used in (for example) parsing programs, because of the great number of variables that play a part in conditioning choices in the text.
The terms climate and weather are related to each other as langue to parole: as the system of language is to its instantiation in text. Weather is the instantiation of climate; climate is the system ‘behind’ the weather. As with language, there is, of course, only one set of phenomena here not two; when we refer to climate we are construing general principles and tendencies that ‘explain’ the multidimensional microvariation that is what we actually have to live with …
Instantiation is the relation between the system and the instance. When we shift attention along this scale, we are moving between the potential that is embodied in any stratum and the deployment of that potential in instances of the same stratum … this move can be made at any degree of delicacy.
Realisation is the relation of one stratum to other strata … when we shift attention ‘upwards’ into context or ‘downwards’ to lexicogrammar and phonology/graphology, we are moving in realisation. We can do this at any degree of delicacy, from the most general to the most specific; and we can do it at any point along the instantiation scale, from system to text.
… delicacy is different from instantiation. Delicacy is the degree of detail, or specificity, to which the description is taken.
Delicacy is the relation between the most general features and the most specific.
Since domains are instantiations of the overall semantic potential, any properties associated with types in this potential are also present in the domains.
On the one hand, registerial domains form families within which there is variation, such as the family of culinary procedures, which encompasses a variety of different kinds of recipes, or the more extended family of procedures for creating artefacts. On the other hand, the variation is multidimensional.
The overall semantic potential is diversified into registerial varieties that emerge as patterns of instantiation.
Register is systematic variation within the overall meaning potential; from this point of view, a domain is a region of registerial variation corresponding to the contextual variable of field — the meaning potential determined by, and determining, ‘what’s going on’.
… when we model a domain, we are still defining a meaning potential, a semantic space in which meanings are constantly being realigned and new meanings created. In other words, we are describing a register; or, more accurately, a region of registerial variation…
Instances do not usually repeat each other word for word; there is some variation … It is this variation that constructs the domain. With each instance of the field, the overall meaning potential is instantiated in slightly varying ways; this variation both confirms the earlier patterns and nudges them along some novel path.
The relationship which links the domain models with the general model is instantiation. Instantiation relates the system to the instance, at any given stratum. … A domain is located somewhere between the system and the instance. Hence it can be looked at from either end. Seen from the ‘system’ end, a domain is regular and repeated patterning in the way potential is deployed, over the course of time. Seen from the ‘instance’ end, it is generalisation across particular instances of text.
Domain models are variants of the general model. A particular domain model specifies which of the semantic systems in the overall model are activated in a particular contextual field: the ideational meanings that are “at risk”. Each field thus has its own semantic profile, which can be seen against the background of the overall semantic potential.
… the semantic correlate of a contextual field is a domain. When we model the ideational semantics of a particular field, we create a domain model.
If the field is defined in relatively broad terms, it may activate most of the general part of the ideation base; but a more restricted field may call on only one particular part of it. In such cases, it may be useful to reduce the whole ideation base to only those parts that are implicated for the particular field; that is, to set up a field-specific ideational semantics. (More generally, this might be a context-specific semantics, including interpersonal and textual as well as ideational meanings). However, the price of this would be to isolate such a field from others to which it is related.
Each field is projected onto one variety of the ideation base: that is, it can be thought of as activating some portion of the total semantic resources. This projection of field onto the ideation base involves both the particular domain and the general types under which this domain is classified. A given field is thus constituted as a principled selection of types from within the ideation base; this is so to speak its semantic image projected from context.
… a full account of field would include a typology of the possible first and second order values that occur in a culture. Such a typology would show how closely various fields are related — how they form families. So, given three different fields, 1, 2, and 3, the typology will show three different ways of deploying the ideation base.
… field can be characterised as the deployment and organisation of the ideation base. … there are two aspects to this category. In most contexts, there is both a first order field and a second order field — the first order field is the social activity being pursued (eg instructing somebody in how to prepare a dish …) and the second order field is the ‘subject matter’ the activity is concerned with (eg the ingredients and methods of cooking …). … And both these guide the way the ideation base is deployed.
… the covariation between context and language is not undifferentiated — it is differentiated according to the functional diversification of each of the two strata: the variables of context correlate respectively with the metafunctions of language, field with ideational, tenor with interpersonal, and mode with textual. These pairs are mutually predictive.
… the notion of functional dialect was worked out by the Prague School in the 1930s … and systemic register theory has its roots in Firth’s work on restricted languages. The concept of register is also recognised in computational linguistics under the heading of sublanguage, particularly in work on machine translation.
Together, field, tenor and mode define the ‘ecological matrix’ in which particular types of text are processed: there is a systematic relationship between such matrices (particular combinations of field, tenor and mode values) and particular types of text. We can see this clearly with actual instances of text — for example, an individual recipe or weather forecast; but these are instances of general classes, to be characterised in terms of the systemic potential that is instantiated in them. That is, there is a correlation not only between a contextual matrix and a given instance of a recipe but also between that matrix and the linguistic potential that is deployed in recipes in general. The latter correlation is known as a functional variety or register of the general systemic potential.
… both these contextual variables [field and tenor] are, in some sense, independent of language, even though they are constituted in language and the other semiotic systems of a culture. That is, they concern realities that exist alongside the reality created by language itself, semiotic reality. However, there is a third contextual variable that is specifically concerned with the part language is playing in any given context — the symbolic mode, how the linguistic resources are deployed. This covers both the medium (spoken, written, and various subtypes such as written in order to be spoken) and the rhetorical function — persuasive, didactic, informative, etc.
The context encompasses both the field of activity and the subject matter with which it is concerned (‘what’s going on, and what is it about?’) and the tenor of the relationship between the interactants, between speaker and listener, in terms of social rôles in general and those created through language in particular (‘who are taking part?). The field is thus the culturally recognised repertoires of social practices and concerns, and the tenor the culturally recognised repertoires of rôle relationships and interactive patterns.
The phenomena of human experience are held in tension by so many intersecting analogical lines that, while all of us have the same brains and live on the surface of the same planet, such diverse ways of semiotic mapping are not only possible but inevitable.
Some process types tend to lie on the borderline between major categories, forming mixed and overlapping categories; typical of these are the behavioural and existential proceses in English. It is likely that equivalent types of process will be liable to greater typological variation than those that fall squarely within the core categories of material, mental, and relational.
… we referred to projection as something that overlaps the ‘boundary’ between interpersonal and ideational metafunctional space; in English it is typically construed ideationally, though with a close relationship to the interpersonal systems of modality and mood. Other languages locate projection rather differently in relation to this boundary, sometimes foregrounding its interpersonal aspects, for example through a special category of ‘reporting’ mood.
We saw that in the experiential grammar of the English clause there was a complementarity of perspective between the transitive and the ergative: processes may be construed either as ‘one participant is doing something, which may or may not extend to another participant that is being done to’, or as ‘one participant is involved in something, which may or may not be brought about by another participant that is the agent of it’. Probably all languages display this transitive/ergative complementarity in their transitivity systems; but at the same time it appears at different depths and in different proportions.
As we have shown in the context of grammatical metaphor, the choice among alternative construals is made on the basis of both ideational and textual factors. These factors ‘conspire’ together so that different strategies are favoured in different registers: the congruent form (sequences) in casual speech, the metaphoric form (figures of being & having) in elaborated forms of writing.
The elasticity of construal is further increased through grammatical metaphor, which also relies on the manifestation of projection & expansion within participants (realised by the different types of modification in nominal groups). We have illustrated the range of options throughout the system at various points, and a schematic example will be enough here:
[i] logicalA happened, so B happened;A caused B to happen;
[ii] experientialA happened causing B / B happened because of AA happening caused B happeningA affected B [‘cause-happened’]
with additional metaphorical variants:B happened because of the happening of Athe happening of A caused the happening of Bthe happening of A was the cause of the happening of B
One major source of this kind of elasticity [construing phenomena as sequence or figure] is the manifestation of the very general semantic types of projection and expansion throughout the ideation base. In English, as we have shown, we find them manifested both logically as relations and experientially as elements within figures:
[i] logical manifestation:
as relations between figures in the construal of sequences (realised by clause complexes);as relations between processes in the construal of figures (realised by verbal group complexes);
[ii] experiential manifestation:as circumstances within figures (realised by prepositional phrases or adverbial groups);
as circumstantial processes within figures of being & having (realised by relational clauses);
as non-nuclear participants within figures, i.e., participants other than Medium — Agent, Beneficiary, Range (realised by nominal groups with or without a preposition).
Given that a process takes place in time, there will be some sense in which it has a beginning, a middle and an end. These may be observable as distinct constituents, if it is a process having duration; even if it is instantaneous, however, they represent possible facets, points of view from which it can be considered. In English, the meaning of a process typically includes it completion: if I ‘cut’ a piece of string, I cut it in two. To construe a process as non-completed, English uses conative or inceptive phase: ‘try to cut it’, ‘start to cut it’.