Rethinking context: realisation, instantiation, and individuation in systemic functional linguistics

3.1 Field as a resource

Following on from Halliday (1978) the basic approach to characterising field in SFL has been a classificatory one.^[5] He (Halliday 1978: 64, 115, 143, 222) notes for example the fields of personal toilet, child at play, games and buying/selling newspapers as goings on. Martin (1992: 544) classifies fields in relation to how they are learned – as domestic, specialised (e.g. sport, craft, trades), administrative (e.g. bureaucracy, legislation, law), or exploratory (science, social science, humanities). Later on, in dialogue with Muller (2000) and Maton (2014), Martin (in Martin and Maton 2017; Martin et al. 2020a) further classifies exploratory fields according to what Muller calls verticality and grammaticality – in simple terms the extent to which fields develop by testing hypotheses in order to establish ever more general understandings covering an ever wider range of data or by developing new ways of interpreting comparable and/or alternative data.

These approaches have proven useful for understanding different spheres of activity. But it has proven difficult to generalise these classifications beyond relatively common-sense descriptions without shifting toward descriptions that impinge upon all three metafunctions. In addition, as Maton (2014) points out, discrete classifications of fields such as this obscures the wide range of variation and contestation inherent in all social fields – no field is everywhere and always the same. Classifications such as those above are often useful to think with, but become difficult to use when confronted with data.^[6]

Accordingly, in this section we will outline field as a resource for construing phenomena. This perspective emphasises the resonance between field and the ideational metafunction in language, as well as offering tools that can link upwards with genre. The model presented here is that of Doran and Martin (2021). Under this model, phenomena can be construed dynamically as a set of activities oriented to some social action or they can be construed statically as relations among items.

Beginning with the static perspective, this involves viewing fields as sets of items that can be arranged into taxonomies. One type of taxonomic relation is that of composition – where items are arranged into a part-whole relations. For example, in a classroom lesson in cell biology, the teacher at one stage reads from the textbook:

Here, the teacher explains that animal cell centrosomes are composed of two centrioles, which in turn include nine groups of three microtubules. We can visualise this as a compositional taxonomy as in Figure 2.

Figure 2:

Compositional taxonomy of a centrosome.

Alternatively, items can be related in terms of type and sub-type into a classification taxonomy. Mukherjee (2022: 67–69) draws on this when he lays out the diversity of cells that occur on Earth:

Every cell on Earth – which is to say every unit of every living being – belongs to one of three entirely distinctive domains, or branches of living organism. The first branch comprises bacteria: single-celled organisms that are surrounded by a cell membrane, lack particular cellular structures found in animal and plant cells, and possess other structures that are unique to them […] We think of them as pathogens – batonella, pneuomococcus, salmonella – because a few of them cause disease […]. We – you and me – inhabit a second branch, or domain, called eukaryotes. The word eukaryote is a technicality: it refers to the idea that our cells, and the cells of animals, fungi, and plants, contain a special structure called a nucleus (karyon, or “kernel” in Greek). This nucleus, as we will soon learn, is a storage site for chromosomes. Bacteria lack nuclei and are called prokaryotes – that is, “before nucleii” […] And now the third branch: archaea. It may be the single most startling fact in the history of taxonomy that this full branch of living beings remained undiscovered until about fifty years ago […] they lack the defining features of the other two domains. (Mukherjee 2022: 67–69)

This stretch of text is organised around three subtypes of organism (in bold above), for which he also gives some examples. We can visualise this classification taxonomy as in Figure 3.

Figure 3:

Classification taxonomy of organisms.

Whereas classification and composition taxonomies are often considered separately from one another, in this excerpt Mukherjee (2022) shows that the classification of types of organism is based upon their composition (at least in this presentation of the subtypes). Eukaryotes are eukaryotes because they have cells with a nucleus; prokaryotes are prokaryotes because they have only one cell, which has a cell membrane; archaea are archaea because they lack the defining features of the other two subtypes. This illustrates that in technical fields, classification and composition and indeed all resources of field are often mutually defining. They together form a web of relations between items that underpin a complex field (what Maton 2014 within Legitimation Code Theory (LCT) calls an epistemological constellation).

A complementary perspective on field is a dynamic perspective that construes phenomena as a set of activities. For example, in our classroom the teacher explains mitosis as an activity where a single cell divides into two daughter cells (underlined).

This example presents mitosis as a single activity – what we can call an unmomented activity. Alternatively, activities can be divided into a series of moments. When Mukherjee (2022: 99) visualises mitosis, he presents a caption that explains it in terms of its moments – what we call a momented activity. After describing that the chromosomes are initially present in loose threadlike forms in the nucleus, he explains:

Then the threads tighten into dense bundles. The nuclear membrane dissolves, and the chromosomes separate into two sides of the cell, as if drawn by some forces. When they’ve fully separated […] the cell splits, generating two new cells. (Mukherjee 2022: 99)

The activities that Mukherjee (2022) describes are as follows (where ^ indicates a sequence):

The threads tighten into dense bundles

^

The nuclear membrane dissolves

^

The chromosomes separate into two sides of the cell, as if drawn by some forces

^

They (the chromosomes) fully separate

^

The cell splits

^

Two new cells are generated

(Mukherjee 2022: 99)

This presents a complementary view to the static one focused on items – it outlines a dynamic unfolding of events. These activities are interconnected with the items and taxonomies that are involved in them. In this instance, the cell splits due to activities associated with two of its parts: the nuclear membrane dissolving and the chromosomes separating.

Finally, both activities and taxonomies can be described in terms of potentially gradable properties. For example, when Mukherjee (2022: 68) describes bacteria, he does so by describing their characteristics (in bold) and the place they live (spatio-temporal properties underlined):

The abundance and resilience of bacteria stagger the mind. Some live in oceanic thermal vents where the water reaches near boiling temperature; they could easily thrive inside a steaming kettle. Some prosper within stomach acid. Yet others live, with seemingly equal ease, in the coldest places on earth, where the land freezes into packed , impenetrable tundra for ten months of the year. They are autonomous, mobile, communicative, and reproductive. (Mukherjee 2022: 68)

As this excerpt shows, these properties can be graded as being higher or lower (known as arrayed), such as the coldest place on earth and near boiling temperature. Although not specified here, arrayed properties are often gauged by being measured or quantified (e.g. 133.61 °F, 100 °C).

Taxonomy, activity and property make up the basic perspectives on field and offer complementary construals of phenomena. But as we have seen, they do not build their meaning in isolation. Their meanings regularly interrelate with each other, as we saw above where the classification taxonomy of types of organism was based upon their differing compositional taxonomies.

In addition, each of these perspectives on phenomena can be reconstrued in terms of the other. For example, the teacher names the activity of a cell dividing into two identical daughter cells as mitosis.

By naming it in this way, the teacher reconstrues the activity of a single cell dividing into two identical daughter cells as an item – mitosis; or what we call an itemised activity. This allows mitosis to function as an item and be related to other items – for exemple by establishing a classification taxonomy of types of mitosis (e.g. orthomitosis, pleuromitosis, extranuclear mitosis, intranuclear mitosis, closed mitosis, open mitosis, etc.). At the same time, mitosis can function as an activity and be sequenced with other activities as, say, moments within the cell cycle, along with interphase, telophase, and cytokinesis. Properties can also be reconstrued – for example when Mukherjee (2022) described bacteria as in terms of their abundance and resilience (thereby reconstruing the properties abundant and resilient as itemised properties).

Reconstruals and interrelations such as these allow for increasingly expansive and integrated construals of phenomena. They provide rich resources to build knowledge of the world around us and to map and explain highly complex fields (Doran and Martin 2021; see also Carr 2023; Chen 2024). Together with the basic relations of field – activity, property, and taxonomy – they also link closely with ideational meanings within language (see Hao 2020), which for SFL theory, allows us to maintain ideational register/metafunction resonance between language and context.

3.2 Tenor as a resource

As with field, the basic approach to characterising tenor in SFL has also been a classificatory one. Halliday (1978: 222) notes for tenor that examples “of role relationships, that would be reflected in the language used, are teacher/pupil, parent/child, child/child in peer group, doctor/patient, customer/salesman, casual acquaintances on a train, and so on”. Eggins and Slade (1997: 52–53) suggest that tenors could be co-classified along four dimensions:

status relations (e.g. customer/salesperson)

affective involvement (e.g. friends/lovers)

contact (e.g. immediate family/one off encounters)

orientation to affiliation (e.g. fellow students/fellow passengers on bus)

(Eggins and Slade 1997: 52–53)

Tenor classification is considerably elaborated by Hasan (2014, 2020: 274, 282, Figures 5 and 7 respectively). Her tenor analysis of a phase of mother/child interaction gives us a glimpse of the delicacy of the co-classification and subclassification involved (Hasan 2020: 297). As specified below, there are four major dimensions of co-classification: agentive role, textual role, social role, and social distance, alongside multiple co-classifications and subclassifications within these systems.

First interactant (child):

agentive role: care receiver;

textual role: speaker/addressee present in ‘mss’ [material situational setting];

social role: hierarchic: lower; invisible; offspring; Australian; female; child;

Second interactant (mother):

agentive role: caregiver;

textual role: speaker/addressee present in ‘mss’;

social role: hierarchic: higher: invisible; mother; Australian; female; mature;

social distance: minimal

As with field, these approaches have offered useful insights into the range of factors that need to be accounted for when considering the sociality of language. But like field, it has proven difficult to push past common-sense classifications such as ‘mother’ or ‘female’ etc. This is especially the case considering the highly contested and variable nature of these demographic distinctions that belie an easy classification as a single feature within context (as both Hasan’s and Cloran’s work have shown; e.g. Cloran 1989; Hasan 2009b).

Perhaps more challenging is how to relate many of these features to those of language – it is not clear how, for example, we are to cleanly relate the categories of ‘mature’ or ‘offspring’ to distinct patterns of realisation within language. Nonetheless if we abstract away from precise categories and consider the dimensions such as contact and status, we are on firmer ground. As Poynton (1990) has shown, variables in terms of contact (how close one is to another person) and status (the degree of ‘equality’ or ‘hierarchy’ in the relationship) can in fact be linked to manifestations of language having to do with how reciprocal the sets of choices are between people (status) and how many meanings can be shared (contact). As we will discuss in Section 3.2 below, these choices are not restricted to the interpersonal metafunction, but in fact impinge upon choices in all metafunctions. Thus while these dimensions can be more clearly associated with patterns of language, we once more face the issue that this approach erodes the metafunction/context resonance upon which SFL’s model of language and context is constructed. Accordingly, we will propose below that these dimensions be interpreted as principles of instantiation, rather than as tenor options within the realisation hierarchy.

Finally, recent work on how people build relationships and community discussed as ‘affiliation’ (Knight 2010; Logi and Zappavigna 2022; Stenglin 2004) and on how people position each other as speakers work toward consensus (Kim et al. 2023; Muntigl 2009; Zhang 2020a, 2020b, 2021) has highlighted the close interaction between interpersonal systems of evaluation (described in Martin and White 2005 in terms of appraisal) and dialogue (described in terms of exchange – Berry 1981a, 1981b; Martin 1992; Ventola 1987 – and speech function – Halliday 1985), as well as parallels between this meaning-making and patterns of alignment within written text (Doran 2020a; White 2020). But there has yet to be a clearly articulated ‘place’ for these interactions in SFL theory.

In response, Doran et al. (2024) propose a new model of tenor as a resource for enacting social relations. This model works to make clear the links between tenor and the interpersonal systems of language, and so maintain SFL’s context/metafunction resonance, as well as providing a map of resources that are used across situations to negotiate social relations and realise distinct genres.

Under this model, people can tender meanings to be engaged with, or they can render meanings that have been put forward. For example in the classroom we have looked at, following the teacher asking the students what is inside the centrosome, a student tenders an answer that the teacher then renders by supporting it (by both repeating the answer and using positive attitude Right, Sweet). This follows a prompt from the teacher “Inside the centrosome is … again?” An arrow shows the direction of the rendering.

Student:	Centrioles	tender
Teacher:	Centrioles. Right. Sweet.	render: support

In other instances, the teacher rejects students’ suggestions. In the following, they do this by using the Modal Adjunct No, plus they’re not condensed yet:

Student:	I thought they’d have to be condensed.	tender
Teacher:	No, they’re not condensed yet.	render: reject

These instances illustrate simple instances where rendering cleanly follows tendering. But in practice, tendering and rendering are often done together. This occurs when in response to what one person puts forward, someone else puts forward something that only implies their support or rejection. An adapted example from our classroom illustrates this when the teacher and a student go back and forth as to what is on a diagram:

Student:	Those are the spindle fibres – the green things.	tender
Teacher:	Those would be your microtubules holding the cell	render: reject	tender
Student:	But they’re moving	tender	render: reject
Teacher:	They’re condensing	render: reject	tender

Here, the student puts forward that the green things are spindle fibres. The teacher rejects this, not by saying explicitly that they are wrong, but by tendering an alternative suggestion (that they are microtubules). However the student is not convinced, and so counters by arguing they’re not moving – both rejecting the teacher’s characterisation and tendering a reason for it – a move which the teacher once more rejects by tendering an alternative position (they’re condensing).

This example is a conflictual one, where the teacher and student are at odds. But in casual conversation, a strategy of tendering and rendering support at the same time forms a crucial means of keeping a conversation going indefinitely (Eggins and Slade 1997) – as it allows both responding to and extending what has been said.

Importantly, these resources do not just occur in dialogue, but allow for different positions to be negotiated in monologue. This typically draws on resources of engagement and attitude, as different positions are put forward, supported, or rejected. Indeed, we saw an example of this above where the teacher said They’re not condensed. Here they drew on the negative not (disclaim: deny in Martin and White’s engagement system 2005), to reject the position that they are condensed.

Mukherjee (2022: 68–69) uses this resource when recounting the discovery of archaea – the third main branch of organisms. In the example below, the positions that are being rendered (supported or rejected) are underlined, and the resources that render them are in italics for support and bold for reject.

In the mid-1970s, Carl Woese, Professor of biology at the University of Illinois at Urbana-Champaign, used comparative genetics – the comparison of genes across various organisms – to deduce that we had mis classified not just some arcane microbe but rather an entire domain of life. For decades, Woese fought a spirited but lonely, bitter war that left him ragged at the edges. Taxonomy wasn’t just missing the point, he insisted, it was missing a whole living domain. Archaea, Woese argued, were not “ almost like ” bacteria or “ almost like ” eukaryotes […].

Many prominent biologists ridiculed or simply ignored Woese’s work […]. But decades later, we have largely accepted, validated, and vindicated his theory, so that archaea are now classified as a distinct, third domain of living creatures.

(Mukherjee 2022: 68–69)

We can lay this out as in Table 1, which indicates the position being tendered, whether it is being supported or rejected, what is showing this and to whom support or rejection is sourced to.

If we rearrange the table in terms of the source and what they support or reject, as in Table 2, we can more clearly see the positions at stake.

This example shows that a wide range of positions can be supported or rejected not just in dialogue, but in monologue as well. In addition, it illustrates that for this negotiation, it is not enough to just focus on attitude or engagement (or for dialogue, exchange); rather one must look at how these resources all together realise general patterns of support or rejection. Importantly, this is done almost entirely through interpersonal systems, allowing us to maintain the context/metafunction resonance between tenor and the interpersonal metafunction.

Although we cannot explore tenor resources in more detail here, the above texts also illustrate how positions can be sourced to different people, and arranged in relation to each other – such as when Mukherjee (2022) draws on but (a resource for countering) to oppose the initial reactions to Woese’s work (sourced to Many prominent biologists, functioning as the Appraiser) to the current feelings (sourced to us):

Many prominent biologists ridiculed or simply ignored Woese’s work […]. But decades later, we have largely accepted, validated, and vindicated his theory.

The ability to both render positions and orient them in relation to people and other positions offers a means of building large networks of meaning that are organised not through ideational configurations (i.e. through field), but through rhetorical relations (what in LCT Maton 2014 calls an axiological constellation). That is, it allows us to map sets of interpersonal values and how they are oriented to different people and communities (Doran 2020a, 2020b, 2024). As noted above, this is done in a way that maintains the connection between tenor and the interpersonal metafunction.^[7]

3.3 Mode as a resource

As with field and tenor, the basic approach to characterising mode in SFL has been a classificatory one. Halliday (1978: 33, 115, 144) mentions speaking/writing, monologue/dialogue, and written to be read aloud to characterise the part language is playing.^[8] Martin (1984, 1992 suggests classifying modes along two scales – one ranging from language in action to language as reflection^[9] and another ranging from dialogue to monologue; in addition he considers (1992: 515) the affordances of different modes as far as visual and aural feedback are concerned – i.e. none, one-way, or two-way. Hasan (2014: 54, 2020: 259) presents a delicate categorisation, built up around her perspective on material contact (means of expression – e.g. graphic/phonic) and semantic contact (e.g. elocution, turn management). As with field and tenor, these approaches offer useful insights into how different media, channels, and relative context-dependence affects language. But as has been shown by Martin and Matruglio (2013) in their work on presence and the expansive work on a range of emerging communicative channels (e.g. Zappavigna’s 2012, 2018 work on social media), as with field and mode, these perspectives impact upon all metafunctions. As such, we prefer to interpret them from the perspective of instantiation and individuation and develop an alternative model of mode – as a resource that resonates with textual meanings.

The model proposed here conceptualises mode as a resource for organising information. As with the models for tenor and field, it does so in a way that aims to make the links between mode and the textual metafunction clear, while at the same time being able to connect with patterns of genre. Under this tentative model, texts can be organised into chunks of information, by demarcating boundaries or sustaining the text’s flow. These options occur within a system called juncture.^[10]

Mukherjee (2022: 68–69) draws on these resources in order to tell the story of the discovery of archaea we saw above. The excerpt we will explore follows reports of two of the branches of life: eukaryotes and prokaryotes. It begins by introducing the third branch archaea at the beginning of a new paragraph by saying:

And now the third branch: archaea (Mukherjee 2022: 68).

Before looking at the whole excerpt, we can note that this sentence functions as a clear demarcation of information from the previous chunks about eukaryotes and prokaryotes. From the perspective of periodicity and theme (Halliday and Matthiessen 2014; Martin and Rose 2007), it puts forward a Hyper-Theme that predicts what is to come (a stretch about archaea) and draws on both a marked and textual Theme (And now) in order to shift the text (Fries 1995). From the perspective of identification (Martin 1992), it both presents a new participant that has not yet previously been mentioned (archaea) and splits an established participant chain (via the third branch) – making clear that there has been a first and second branch, and that this is a separate third branch. And finally, from the perspective of connexion (Halliday and Hasan 1976; Martin 1992), the sentence-initial internal addition conjunction And signals a new stretch of language being connected to the previous. Three sets of textually oriented systems, then – periodicity, identification, and internal connexion – work together to indicate that there is a new chunk of information coming up.

This sentence introduces a longer chunk of information:

And now the third branch: archaea . It may be the singularly most startling fact in the history of taxonomy that this full branch of living beings remain undiscovered until about fifty years ago. In the mid-1970s, Carl Woese, Professor of biology at the University of Illinois at Urbana-Champaign, used comparative genetics – the comparison of genes across various organisms – to deduce that we had misclassified not just some arcane microbe but rather an entire domain of life. For decades, Woese fought a spirited but lonely, bitter war that left him ragged at the edges. Taxonomy wasn’t just missing the point, he insisted, it was missing a whole living domain. Archaea, Woese argued, were not “almost like” bacteria or “almost like” eukaryote. (“Almost-like” is the taxonomist’s version of a parent saying to a child, “Go away, you’re bothering me.”)

Many prominent biologists ridiculed or simply ignored Woese’s work. In 1998, Ernsts Mayr, the biologist wrote an essay on Woese drenched with teacherly condensation (“Evolution is an affair of phenotypes […] not genes.”), getting the story exactly wrong. It wasn’t evolution that Woese was contesting, it was taxonomy – which is precisely the question of genes. A bat and a bird may have nearly the same physical characteristics, or phenotypes. It’s the differences in their genes that gives away the secret: they belong to different taxa. The journal Science described Woese as a “scarred revolutionary.” But decades later, we have largely accepted, validated, and vindicated his theory, so that archaea are now classified as a distinct, third domain of living creatures.

(Mukherjee 2022: 68–69)

As we noted above, the opening sentence predicts that the following text will talk about archaea. This is done through participant tracking: the entity archaea is presented, and then referred to (in bold above) as this full branch of living beings and as archaea twice. In addition this entity is classified as not just some arcane microbe, an entire domain of life, and a whole living domain. The final reference to archaea occurs in the final clause archaea are now classified as a distinct, third domain of living creatures, that functions as a Hyper-New synthesising the point of the excerpt. This Hyper-New demarcates the boundary between this excerpt and the following text. But between the opening and closing sentences, the participant chain and classification (amongst other things), functions to sustain the chunk of information and make clear that we are still focusing in some sense on the same stretch of information.

Within this excerpt, there are other smaller demarcations. For example, in the first paragraph, the marked Theme In the mid-1970s, indicates that there will be a shift – in this case to a story about Carl Woese (who is introduced in this clause). The first sentence as a whole functions as a smaller Hyper-Theme for this story:

In the mid-1970s, Carl Woese, Professor of biology at the University of Illinois at Urbana-Champaign, used comparative genetics – the comparison of genes across various organisms – to deduce that we had misclassified not just some arcane microbe but rather an entire domain of life.

For decades, Woese fought a spirited but lonely, bitter war that left him ragged at the edges. Taxonomy wasn’t just missing the point, he insisted, it was missing a whole living domain. Archaea, Woese argued, were not “almost like” bacteria or “almost like” eukaryote. (“Almost-like” is the taxonomist’s version of a parent saying to a child, “Go away, you’re bothering me.”)

Many prominent biologists ridiculed or simply ignored Woese’s work.

In 1998, Ernsts Mayr, the biologist wrote an essay on Woese drenched with teacherly condensation (“Evolution is an affair of phenotypes […] not genes.”), getting the story exactly wrong. It wasn’t evolution that Woese was contesting, it was taxonomy – which is precisely the question of genes. A bat and a bird may have nearly the same physical characteristics, or phenotypes. It’s the differences in their genes that gives away the secret: they belong to different taxa. The journal Science described Woese as a “scarred revolutionary.”

But decades later, we have largely accepted, validated, and vindicated his theory, so that archaea are now classified as a distinct, third domain of living creatures.

(Mukherjee 2022: 68–69)

This first paragraph tracks Woese’s argument, focusing primarily on Woese himself. The paragraph break then leads to a new Hyper-Theme that introduces other biologists’ perspectives (Many prominent biologists ridiculed or simply ignored Woese’s work) – which involves establishing a new participant chain involving countering authorities (Many prominent biologists, Ernsts Mayr, the biologist, The journal Science). Put in terms of the model of mode we are introducing here, each of the opening lines – starting with In the mid-1970s, For decades, and Many prominent biologists – all function to demarcate distinct chunks of information. But they do this within the larger chunk we have already established that focuses on archaea more broadly. This illustrates that there is not just a single linear chunking of information in a text, but rather that texts can be organised around a hierarchy of demarcation – whereby smaller chunks of information are organised within larger chunks of information.

Indeed if we zoom out to an even longer stretch than we have here, the whole excerpt functions together with the previous sections on eukaryotes and prokaryotes as a single, larger chunk introducing what organisms are. This is suggested by the internal addition connexion And (which suggests a linking with the previous chunk) and the chain splitting the third branch (which, while introducing archaea and in doing so establishes a new participant chain) which in fact refer back to and illustrate similarity with the previous two branches. Importantly, this hierarchy of demarcation is developed not by a single discourse semantic system, such as periodicity, identification, or internal connexion,^[11] but by them all working together. And since these systems are all primarily concerned with organising texture, the resonance between textual metafunction and mode is sustained. Establishing hierarchies of demarcation also conforms to the general periodic structure of textual systems, whereby the same ‘meanings’ can be overlayed on top of each other at bigger of smaller stretches (Halliday 1979).

Below, the excerpt is replayed with boxes indicating some of the demarcation, with smaller chunks nested within larger ones.^[12]

In addition to chunking up the text, mode also functions as a resource for foregrounding and backgrounding information, through a system called pulsing. In the excerpt above, this is most clearly realised through the Hyper-Themes and Hyper-News which foreground the key meanings that occur. In this case, what is foregrounded is the ideational category of archaea as a branch of life in the Hyper-Theme and the eventual interpersonal vindication of this category and Woese in the Hyper-New:

And now the third branch: archaea.

[…]

But decades later, we have largely accepted, validated, and vindicated his theory, so that archaea are now classified as a distinct, third domain of living creatures.

Foregrounding of this kind is potentially realised through a range of systems including periodicity in discourse semantics, theme in lexicogrammar, information and salience in phonology, as well as paralinguistic vocal features of sound quality, including loudness, tenseness, pitch height (van Leeuwen 1999), the use of gestural beating (Ngo et al. 2022b) and salience within visual multimodal texts (Kress and van Leeuwen 2020). It offers a resource for composing texts in terms of pulses of prominence, foregrounding, and backgrounding information as a text flows.

Finally, mode offers resources for distributing information across a text, via a system called distribution. The basic distinction is between an instance of language that indicates that there is further information needed to understand what is being said (i.e. that the information is in some sense distributed) or that all the information is given in the immediate instance.

Distribution of information occurs throughout the classroom example we have been looking at, where the teacher refers across modalities to the slides they are using (in bold below):

I have this image here of the cell undergoing mitosis for two replicated daughter cells. You’ve got here DNA replication with the cell cycle – what part is that called?

In this example, the teacher is specifying that the information needed is distributed between the spoken language and the slide they are looking at. She does this by drawing on exophoric reference to the infographic (Halliday and Hasan 1976; Martin 1992). This contrasts with a number of instances at the beginning of the class, where the teacher is establishing what they are going to do. In these instances, marked in bold, the teacher is indicating that there is no relevant information elsewhere that is needed to follow what is going on – the information is not distributed but immediate.

What we’re going to do today is model a representation of mitosis as a point of reference to explain these stages. I’m going to give you some materials so when we jump into our groups you’re going to construct a model for the stages.

The immediacy of the information is established in three of these instances through presenting reference (Martin 1992) – a representation of mitosis, some materials, a model, where the indefinite Deictics indicate that these participants are being introduced and are not to be recovered from elsewhere in the text or the situation. The fourth possible instance, our groups, draws on homophora (Halliday and Hasan 1976); in this case the entity’s identity is presumed, and so does not need to be recovered from anywhere else.

Distribution can also occur within texts. The clearest instance of this involves drawing on anaphora – whereby an instance refers backwards in a text, often via pronouns. In the following text, the teacher first puts forward the entity ‘nucleolus’, and then each subsequent reference uses it to indicate the information being sought is distributed across the text:

Phoricity resources such as those noted above are one of the key means of distributing information. But they are by no means the only resource for doing this. In the text above, the two questions put forward by the teacher and student: Do you know what the nucleolus is? And Isn’t that where it make ribosomes? also make clear that the information of the text is to be distributed – in this case that another move is needed for closure. Similarly, the use of text reference (Martin 1992), often coupled with semiotic entities (Hao 2020) and connexion resources, also indicates a distribution of information. In the following (constructed) example, the semiotic entity three main arguments points forward to the fact that a set of arguments are required for the full meaning of the section; the internal connexions First, Second, and Finally signal that each argument connects either forward or backward with the other arguments; and the text reference None of these arguments at the end, looks backwards by making clear that the information that is ‘not convincing’ is to be found earlier in the text. These resources all work together to distribute information throughout the text (while also working to demarcate and foreground different components of this information).

There have been three main arguments against students wearing uniforms. First, they dampens students’ individual expression. Second, uniforms are expensive. And third, they harken back to an old-fashioned time of rigid uniformity. None of these arguments are convincing.

As for tenor and field, these systems of mode – distribution, juncture, and pulsing – offer a model of the resources drawn on to organise text, rather than a common-sense classification of different modes. This model of mode as a resource is considerably less developed than that for field (Doran and Martin 2021) and for tenor (Doran et al. 2024), but it nonetheless offers a means of maintaining the context-metafunction resonance that has underpinned SFL’s conception of the relationship between the internal and external functionality of language. Of course, this does not mean that other considerations often grouped under field, tenor, and mode (e.g. degrees of technicality, social contact, and context-dependence) do not need to be accounted for. Rather, it means that they need to be conceptualised in a theoretically clearer manner. To do this, we propose a perspective from instantiation that treats these and other variables as coupling principles – i.e. as principles for the co-selection and arrangement of choices in language.

4.1 Mass

Responding to Maton’s characterisation of semantic density as involving ‘formal definitions, empirical descriptions or feelings, political sensibilities, taste, values, morals, affiliations’, we have revisited work on technicality in an effort to broaden SFL’s conception of specialised knowledge. In terms of SFL’s concept of metafunction this means extending the focus on ideational meaning to include interpersonal and textual perspectives as well. As noted above, taken together, the contributions from the different metafunctions are referred to as mass (introduced in Martin [2017] and further elaborated with respect to the analysis of infographics in Martin and Unsworth [2024]).

4.2 Association (status/contact)

The tenor specialist whose work most strongly reflects a concern with patterns of usage is Poynton (1984, 1985, 1990, 1996. In her PhD thesis (1990) she refers to these patterns as reciprocity, proliferation, and contraction. Reciprocity refers to the tendency for interlocutors with equal status to take up the same kind of linguistic choices and for interlocutors with unequal status to draw on different ones (the influence of Brown and Gilman’s [1960] canonical study of pronouns is apparent here). Poynton’s research focused on naming and highlights the contrast between reciprocal usage (e.g. both using first names: Kate/Gunther) and non-reciprocal usage (e.g. one using title + surname and the other using first name: Dr. Smith/Gunther). Proliferation refers to the tendency for interlocutors to take up a wider range of choices the better they know one another (e.g. the names that a distant work acquaintance may use for one of the authors, e.g. Dr. Yaegan Doran, vs. the names his mother uses for him, e.g. Yaegan, Yaegan John, Yaegan John Doran, Yaeg, Yaegy J, Yaegy JJ, Yaegy JJJ, Yaegy JJJJJ, B, BB, my B, my BBBBB – with all the nuance in warmth or warning that each choice entails). Poynton also notes the concomitant tendency to prefer more contracted realisations (e.g. Yaeg vs. Dr. Yaegan Doran) the more contact we have with one another.

Martin (1992: 529, 532) provides more detail on the language systems in play as far as Poynton’s patterns of usage are concerned, concentrating on interpersonal systems. But even there textual systems (e.g. rhythm, homophora) and ideational systems (e.g. agency, technical lexis) are brought in to provide a more complete picture. This suggests that as far as social relations are concerned, a multi-functional perspective will prove useful. Below we suggest a framework for what we will refer to as association – comprising ideational (participation), interpersonal (accord), and textual (coordination) patterns of usage as interlocutors enact relations of status and contact with one another.

4.3 Presence

Responding to Maton’s (2014: 110) characterisation of context-dependence as involving “processes of strengthening semantic gravity, such as moving from abstract or generalized ideas towards concrete and delimited cases, and weakening semantic gravity, such as moving from the concrete particulars of a specific case towards generalizations and abstractions whose meanings are less dependent on context”, we have revisited work on context dependency in SFL to include ideational and interpersonal perspectives alongside the more traditional textual ones. As noted above, taken together, the contributions from the different metafunctions are referred to as presence (Martin and Matruglio [2013] who draw on earlier work on context dependency by Martin [1983, 1984, 1992] and Cloran [1994, 1999a, 1999b, 2000]). This problematizes concerns in SFL for the cline between action and reflection (e.g. Martin [1984] and related concerns in the work of Hasan [1999] and Cloran [1994]), and monologue versus dialogue, acknowledging that they do not simply impinge upon textual meanings, but meanings across all metafunctions.