Misrepresenting The SFL Orientation To Meaning

Martin, Matthiessen & Painter (2010: 171):

When we are first working our way into the linguistic system as pre-school children, nominal groups are used to construe the concrete sensible objects of our experience of the material world. They refer to the people and things that we can see, hear, smell, taste and feel.

 

Blogger Comments:

This is potentially misleading, because it invites a transcendent view of meaning, wherein meaning transcends semiotic systems, and the meanings of semiotic systems refer to the meanings outside semiotic systems. This is the opposite perspective on meaning to the immanent perspective on which SFL Theory is founded, wherein meaning is exclusive to semiotic systems; see, for example, Halliday & Matthiessen (1999: 415-8). 

In this view, experience is construed as meaning, with the meanings of perceptual semiotic systems reconstrued as the meanings of language that are realised by the wordings of language, such as nominal groups. That is, in SFL Theory, nominal groups do not refer; they realise the meanings (people and things) construed of experience.

A Problem With A Further Proposed 'Classifying' Type Of Focus

 Martin, Matthiessen & Painter (2010: 170-1):

Closely related to these ‘of’ expansions are groups that would be considered appositive in traditional terms. Here there is a relation between a thing and its class identity: the month of May, the state of Alabama, the city of Rome, the kingdom of Norway, the house of Windsor, the science of cladistics, the title of Mayor (cf. the month is May, the state is Alabama, etc.). We could assimilate these with the classifying type above.

Blogger Comments:

To be clear, in the 'classifying' subtype of Focus, the logical Head of the nominal group is realised by various words for 'subtype' (see previous post). Here, in contrast, the Head is not realised by a word for subtype, but by specific types of superordinate. For example, 'month' is neither a word for 'subtype', nor a subtype of 'May'. Consequently, even ignoring the problems outlined in the previous post, "assimilating" these with the classifying type of Focus is inconsistent within the authors' own model.

In short, in terms of SFL Theory, these are simply nominal groups in which Head conflates with Thing:

Serious Problems That Invalidate The 'Classifying' Subtype Of Focus

   Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

Blogger Comments:

To be clear, contrary to the above, the 'classifying' subtype of Focus does not classify a sub-type of the Thing. Instead, it simply construes the notion of 'subtype'. 

In the second two editions of IFG (Halliday & Matthiessen 2004; 2014: 395), this type of construction is interpreted as an extended Numerative:

The problem here is that 'type (quality)' is inconsistent with the category meaning of 'Numerative'. Halliday (1985: 163; 1994: 183):

The Numerative element indicates some numerical feature of the subset: either quantity or order, either exact or inexact.

On the other hand, in the first two editions of IFG (Halliday 1985; 1994), this type of construction is not identified as one in which Head and Thing are out of phase, so it is simply a case of Head/Thing conflation:

This interpretation is supported by the fact the Head is also the entity that functions as participant in the transitivity system (Halliday & Matthiessen 2014: 394). For example, it is the Head element that is also the Thing to which Attributes are assigned:

This kind of X is tasty.

This type of X is unreliable.

This class of X is superior.

This category of X is illegal.

This brand of X is over-priced.

This make of X is downmarket.

This form of X is beautiful.

This variety of X is popular.

This species of X is endangered.

The 'Evaluative' Subtype Of Focus Reconsidered

  Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

Blogger Comments:

To be clear, the 'evaluative' subtype of Focus is the attitudinal counterpart of the Epithets in the 'selecting' subtype of Focus. On the basis of the same reasoning that Halliday used to derive 'Pre-Numerative' and 'Pre-Deictic', these can again be interpreted as 'Pre-Epithets'. Halliday & Matthiessen (2014: 394):

When the Head is conflated with the Epithet or the Classifier, the relationship between the Head and the Thing is an elaborating one (which is why this type of construction has been called appositive): ‘Head is Thing’.

 

cf.

A Serious Problem That Invalidates The 'Dimensional' Subtype Of Focus

 Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

Blogger Comments:

To be clear, not one of examples of the 'dimensional' subtype of Focus is a nominal group with the Head and Thing out of phase. That is, in every single case, the Head conflates with Thing:

This can be demonstrated by the fact the Head is also the entity that functions as participant in the transitivity system (Halliday & Matthiessen 2014: 394). For example, the Head element is the Thing to which Attributes are assigned:

The size of X is enormous.

The shape of X is irregular.

The colour of X is crimson.

The height of X is 2 m.

The weight of X is 4 kg.

The length of X is 40 km.

The texture of X is silky smooth.

The taste of X is like syrup.

The smell of X is like coal gas.

The nose of X is floral.

The aroma of X is fragrant.

The feel of X is unpleasant.

The look of X is modern.

The sound of X is mellifluous.

The 'Selecting' Subtype Of Focus As A Confusion Of Numeratives And Epithets

 Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

⁴ Some selecting Focus items appear without the structure marker of (e.g. half the boys, all the problems).

Blogger Comments:

To be clear, from the perspective of SFL Theory, the 'selecting' subtype of Focus includes not just quantitative (three, four, some, none, hundreds, thousands) and ordinative (first, third, last, next) Numeratives but also Epithets (bigger, smaller, largest, smallest).

The Numeratives in the 'selecting' subtype of Focus are just Pre-Numeratives on the SFL model:

In such cases, the fact that numeral serves as Head is shown by 'subject-verb agreement': one of the awards is… vs three of the awards are… . Halliday & Matthiessen (2014: 394):

It is the Head that determines the value of the entity in the mood system, and therefore as a potential Subject.

On the basis of the same reasoning that Halliday used to derive 'Pre-Numerative' and 'Pre-Deictic', these Epithets can be interpreted as 'Pre-Epithets'. Halliday & Matthiessen (2014: 394):

When the Head is conflated with the Epithet or the Classifier, the relationship between the Head and the Thing is an elaborating one (which is why this type of construction has been called appositive): ‘Head is Thing’.

cf.

Rebranding Halliday's 'Complex' Numeratives As The Authors' 'Re-counting' & 'Partitive' Subtypes Of Focus

 Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

Blogger Comments:

To be clear, the 're-counting' and 'partitive' subtypes of Focus are merely the authors' rebranding of Halliday's types of complex Numerative. Halliday (1985: 173-4; 1994: 195):

The 're-counting' subtype of Focus includes both quantitative Numeratives (cup, glass, jug, tank, pound, yard, metre, gallon, litre) and collective Numeratives (set, collection, group, crowd, school, herd, flock, gaggle, pod, swarm).

The 'partitive' subtype of Focus is just a rebranding of partitive Numeratives.

Problems With The 'Perspective' Subtype Of Focus

Martin, Matthiessen & Painter (2010: 170):

The Focus element can have a range of functions, which are outlined in Table 5.7 below.

Blogger Comments:

To be clear, the Focus subtype 'perspective' combines nominal groups with Pre-Deictics with nominal groups with Head and Thing conflated. 

The Pre-Deictic expressions are side, top, front, back, foot, peak, summit, root, start, end:

The Head/Thing nominal groups feature picture, painting, photo, drawing:

To be clear, in such nominal groups, picture, painting, photo, drawing is the entity that functions as the participant in the transitivity structure, and so, as Thing as well as Head. Halliday & Matthiessen (2014: 394):

What all these [out-of-phase constructions] have in common is that, while the Thing (e.g. coffee) is the entity that is functioning as participant in the transitivity structure of the clause, the logical Head of the construction is something that constrains the entity in terms of the two variables mentioned above. It is the Head that determines the value of the entity in the mood system, and therefore as a potential Subject.

For example, in the clause Oscar burnt a picture of Dorian, the thing that was burnt was the picture, not Dorian.

Needlessly Rebranding Halliday's 'Numerative' As The Authors' 'Focus'

 Martin, Matthiessen & Painter (2010: 170): 

Nominal group expansions of this kind are very common, and we’ll suggest an analysis here that treats these structures as having an embedded nominal group with the multivariate function Focus; in these nominal groups of functions as a structure marker, signalling the embedding (there is no need to label it separately in the analysis).

 

Blogger Comments:

To be clear, on the one hand, this is an unnecessary rebranding of the term Numerative; Halliday (1994: 195): 

On the other hand, the term 'Focus' is a very poor choice, because the term is already in use in the theory (the Focus of New information), and in this usage, the term does not construe experiential meaning. Moreover, it creates unnecessary confusion because the unmarked Focus of such a nominal group is on the other element, the Thing:

Not Acknowledging Or Distinguishing Pre-Deictics And Pre-Numeratives

Martin, Matthiessen & Painter (2010: 169-70):

However, there is one situation where the two types of structure are out of phase with one another even though the nominal group is a full one (see IFG2: 195-196, IFG3: 332-5 for grammatical indications of the two structures being out of phase). This happens when the Head is a noun that does not represent a thing in its own right but rather an elaboration or extension of another thing, as in the side of the house, two metres of fabric, another cup of coffee. In such cases the multivariate and univariate structures are out of phase with one another:

Here the univariate Head has a kind of ‘numerative’ function, but in fact, the full potential of the nominal group (including Deictic, Post-Deictic and Numerative) is still available after this initial Numerative choice. We might have, for example, two metres of the first fabric [[I looked at]].

Blogger Comments:

[1] This is potentially misleading. IFG2 (Halliday 1994: 195-6) makes a distinction between structures exemplified by the side of the house versus two metres of fabric. The former type involves a Pre-Deictic element, while it is only the latter type that involves a Pre-Numerative:

 

Note that, in Matthiessen's editions of IFG (2004: 333; 2014: 395), this distinction is lost, as both types are interpreted as extended Numeratives, despite the fact that Pre-Deictics are types (qualities), not measures (quantities).

[2] This is potentially misleading, because this fact is already acknowledged in IFG2 (Halliday 1994: 195):

In the experiential structure, therefore, it is the Numerative that is embedded; and since it is embedded, it comes to the front and may be followed by a fully structured nominal even beginning with a Deictic, as in a cup of that good strong tea.

and it is this that motivates the terms 'Pre-Numerative' and 'Pre-Deictic'.

Misrepresenting Conflation

Martin, Matthiessen & Painter (2010: 169):

2.1.3 Conflation of multivariate structure and univariate structure

Blogger Comments:

This misrepresents conflation, because it misunderstands structure. It is elements of structures that can be conflated, not structures. This is because a structure consists of the relationships between elements. Halliday & Matthiessen (2014: 451):

Note that, although it is the functions that are labelled, the structure actually consists of the relationships among them.

and relationships of different metafunctions cannot be conflated. For example, the relationship of the Subject to the Adjunct in the clause they were defeated by the referee cannot be conflated with the relationship of the Medium to the Agent.

Misrepresenting New Information

Martin, Matthiessen & Painter (2010: 169):

Qualifiers are typically realised by ‘downranked’ units – embedded clauses or prepositional phrases, and this is one reason why they occur after rather than before the Thing. Another reason derives from the textual management of the ‘flow of information’ within the nominal group. As we noted above, the Deictic is a kind of nominal group Theme analogous to a clause Theme. Similarly, the other end of the nominal group, where the Qualifier occurs, has the potential for providing new information just as the end of the clause is the location for the New.

Blogger Comments:

This is misleading. The end of the clause is the unmarked location of New information, in the case of unmarked information distribution (coterminous clause and information unit). To be clear, information can be highlighted as New anywhere in the clause, and information units can be less or more extensive than a single clause.

If this seems a trivial point, Martin's model of New information  — partially a misunderstanding of Fries' 'point' — depends on this false claim. Martin (1992: 452, 454):

Taking New as the final clause constituent in [6:36] displays the same difference in the range of realisations noted for the spoken text considered above. The Theme grounds the genre, anchoring it to just a few meanings and the News articulate the field …

This complementarity of hyper-Theme predicting a text's method of development and hyper-New accumulating its point is outlined in Fig. 6.11.

Misrepresenting The Nominal Group 'Numerative'

 Martin, Matthiessen & Painter (2010: 168-9):

Similarly, Numeratives are concerned with the numerative properties of particular referents (their order in a series or the size of the set of referents) rather than with general experiential classes.

Blogger Comments:

To be clear, here again the authors mistake the Thing of a nominal group for the referent signalled by a reference item. Halliday & Matthiessen (2014: 374) explain the function of the Numerative as follows:

The Numerative element indicates some numerical feature of the particular subset of the Thing: either quantity or order, either exact or inexact.

Seriously Misrepresenting The Nominal Group 'Postdeictic'

Martin, Matthiessen & Painter (2010: 168):

In contrast, at the other end of the cline, the Postdeictic is concerned with particular referents; it serves to indicate how particular referents are selected from some general class – for example, particular referents that are the same or different from another set of referents (the same sedimentary rocks, some different sedimentary rocks). 

Blogger Comments:

This continues the authors' confusion of deixis and reference. Having previously confused the Deictic function in nominal group structure with the cohesive function of demonstrative reference — see previous post — here they confuse the Postdeictic function in nominal group structure with the cohesive function of comparative reference. And again, they confuse the referent (that is referred to) with the nominal group featuring the reference item (that does the "referring).

So, to be clear, the Postdeictic is not concerned with particular "referents"; instead, it identifies a subset of the Thing of the nominal group, as Halliday (1994: 183) and Halliday & Matthiessen (2014: 373) explain:

The post-Deictic identifies a subset of the class of ‘thing’ by referring to its fame or familiarity, its status in the text, or its similarity/dissimilarity to some other designated subset.

So the Postdeictic in the same sedimentary rocks and some different sedimentary rocks identifies a subset of the class of sedimentary rocks.

Importantly, the Postdeictic function is distinct from the comparative reference function of such words, which Halliday & Matthiessen (2014: 632, 633) explain as follows:

In comparative reference, the reference item still signals ‘you know which’; not because the same entity is being referred to over again but rather because there is a frame of reference – something by reference to which what I am now talking about is the same or different, like or unlike, equal or unequal, more or less. Comparative reference items function in nominal and adverbial groups; and the comparison is made with reference either to general features of identity, similarity and difference or to particular features of quality and quantity.

Any expression such as the same, another, similar, different, as big, bigger, less big, and related adverbs such as likewise, differently, equally, presumes some standard of reference in the preceding text.

So the comparative reference items in the same sedimentary rocks and some different sedimentary rocks presume some standard of reference in the preceding text.

Seriously Misrepresenting Nominal Group Deixis

Martin, Matthiessen & Painter (2010: 167):

The Deictic provides the key to the textual status of the particular referent presented by the nominal group. For example, his dog, that dog, the dog are all shown to be identifiable or recoverable by the hearer, whereas a dog, some dogs are shown to be non-recoverable. The Deictic thus ‘contextualises’ the nominal group (compare the way a Theme ‘contextualises’ a clause). The choices associated with the Deictic all concern particular referents at a given point in the discourse – their recoverability status, their location relative to the speaker, etc. At the same time, the Deictic also embodies the interpersonal orientation of the group: it indicates whether the group gives information (e.g. this dog, a dog) or demands information (e.g. which dog?).

Blogger Comments:

[1] To be clear, this misrepresents the function of the Deictic as an element of nominal group structure. As Halliday (1985: 160; 1994: 181) explains:

The Deictic element indicates whether or not some specific subset of the Thing is intended; and if so, which. It is either (i) specific or (ii) non-specific.

This is the nominal group system of DETERMINATION (Halliday & Matthiessen 2014: 365, 366):

[2] Here again the authors mistake nominal groups that either feature a reference item (his, that, the), or don't, for referents. To be clear, the referent is that which is referred to by his, that, the. And again, the authors mistake the recoverable identity signalled by the reference item (e.g. the person signalled by his) for the identifiability (of the Thing) of nominal group featuring the reference item (his dog).

Importantly, reference is not a system of the nominal group (or any other grammatical unit) because it is not a relation that is realised through nominal group structure. The recoverable identity of a reference item lies outside the nominal group featuring the reference item.

[3] This is very misleading. It is not the reference item that 'contextualises the nominal group', as Theme 'contextualises the clause', but the Deictic element as specified by the system of DETERMINATION (see [1] above). Halliday (1985: 166; 1994: 187) and Halliday & Matthiessen (2014: 381):

So there is a progression in the nominal group from the kind of element that has the greatest specifying potential to that which has the least; and this is the principle of ordering that we have already recognised in the clause. In the clause, the Theme comes first. We begin by establishing relevance: stating what it is that we are using to introduce this clause into the discourse, as ‘this is where I’m starting from’ – typically, though by no means necessarily, something that is already ‘given’ in the context. In the nominal group, we begin with the Deictic: ‘first I’ll tell you which I mean’, your, these, any, a, etc. So the principle that puts the Theme first in the clause is the same as that which puts the Deictic first in the nominal group: start by locating the Thing in relation to the here-&-now – in the space-time context of the ongoing speech event.

[4] To be clear, this is indeed the system of DETERMINATION, not REFERENCE (see [1] above).

Misrepresenting Systemic Features As Structural Elements

 Martin, Matthiessen & Painter (2010: 166):

Together, the experiential, interpersonal and textual metafunctions determine the multivariate structure of the nominal group. For example, the nominal group an absolute fool in I think she’s made an absolute fool of herself, has the multivariate structure shown in Figure 5.1.

 

 

Blogger Comments:

To be clear, Figure 5.1 misrepresents systemic features — non-specific, non-interrogative, "interpersonal boost", negative connotation, conscious being — as structural elements. In SFL Theory, only the Deictic, Epithet and Thing constitute the multivariate structure of this nominal group.

Halliday (1994: 190-1) explains how the different metafunctions are embodied in the multivariate structure of the nominal group:

We noted at the beginning of the chapter that in analysing group structure it is not necessary to set up three distinct ‘lines’ corresponding to the experiential, interpersonal and textual metafunctions. A single structural representation will suffice.

We have been able to express this in experiential terms, because it is a general principle of linguistic structure that it is the experiential meaning that most clearly defines constituents. Interpersonal meanings tend to be scattered prosodically throughout the unit; while textual meanings tend to be realised by the order in which things occur, and especially by placing of boundaries. These are very general tendencies, worked out differently in every language but probably discernible in all. …

Since we are using particle theory (constituency) as the foundation of the present analysis — it tends to be conceptually and operationally simpler than models of wave or field — it is natural to represent the structure of the nominal group, in which the functional components are (in English) rather clearly defined, in straightforwardly experiential terms. We shall say little more about the other components, beyond recognising their presence in what has already been discussed. 

(1) Interpersonal meanings are embodied (a) in the person system, both as pronouns (person as Thing, e.g. she, you) and as possessive determiners (person as Deictic, e.g. her, your); (b) in the attitudinal type of Epithet, e.g. splendid; (c) in connotative meanings of lexical items functioning in the group, and (d) in prosodic features such as swear-words and voice quality.

(2) Textual meaning is embodied throughout the entire structure, since it determines the order in which the elements are arranged, as well as patterns of information structure just as in the clause (note for example that the unmarked focus of information in a nominal group is on the word that comes last, not the word that functions as Thing. 

Seriously Misrepresenting The Embodiment Of The Metafunctions In The Nominal Group

Martin, Matthiessen & Painter (2010: 166):

Like a clause, a nominal group embodies all three metafunctions.  

Experientially a nominal group construes a participant – a conscious being (human or animal), an animal not treated as conscious, an institution, a discrete object, a substance, or an abstraction. The group provides the resources for construing a participant as a thing located somewhere in a taxonomy of things expanded by a range of different qualities.  

Interpersonally, a nominal group enacts a person – a role defined by reference to the interaction as either an interactant (‘first/second person’) or a non-interactant (‘third person’) and it can be expanded by various interpersonal assessments.  

Textually, a nominal group presents a discourse referent – a participant and person created as part of a message with status as recoverable (identifiable) to the addressee (e.g. that bird) or as non-recoverable (non-identifiable) at a given point in the discourse (e.g. a bird).

 

Blogger Comments:

[1] This is misleading. A clause embodies all three metafunctions in the systems and structures of the clause. So, like a clause, a nominal group embodies all three metafunctions in the systems and structures of the nominal group. (See the following post.)

However, the authors do not address the question of how all three metafunctions are embodied in the systems and structures of the nominal group. Instead, they focus on the semantics that the nominal group, as a unit, realises.

[2] To be clear, this is the semantics of 'participant', the ideational element that the nominal group realises. That is, the focus is on semantics, not grammar.

[3] To be clear, this is the semantics of a proposed interpersonal element that the nominal group is proposed to realise. That is, the focus again is on semantics, not grammar. But consider the so-called 'persons' enacted in:

[4] To be clear, this is the semantics of a proposed textual element that the nominal group is proposed to realise. That is, the focus again is on semantics, not grammar. But consider the so-called 'referents' presented in:

Most importantly, neither that bird nor a bird is a referent unless referred to elsewhere in a text. That is, the authors mistake a nominal group with a reference item (that bird) — or without (a bird) — for a referent. Moreover, the authors mistake the recoverable identity signalled by the reference item (that) for the identifiability (of the Thing) of the nominal group featuring the reference item (that bird).

These types of misunderstandings of cohesive reference can be found in Martin (1992), Matthiessen (1995), and the two editions of IFG edited by Matthiessen (2004, 2014). For a a self-consistent exposition of reference, see the original model in Halliday & Hasan (1976), Halliday (1985) and Halliday (1994).

Confusing Formal Constituency And Function Structure

Martin, Matthiessen & Painter (2010: 164-5):

Further, multivariate structures are all modelled as constituency structures – as configurations of component parts of a unit whole (see IFG2 Section 2.1, IFG3 Section 1.1.3 on grammatical constituency).

There are thus two basic types of grammatical structure whose characteristics are summarised in Table 5.4.

Blogger Comments:

[1] This is potentially misleading. On the one hand, Halliday & Matthiessen (2004: 384, 383) explain:

Although we have modelled all multivariate structures in terms of constituency, this structural mode is in fact most appropriate for the experiential mode of meaning. That is, the relationships among the elements in a multivariate structure can be characterised as segmental from an experiential point of view but as prosodic from an interpersonal point of view and as culminative from a textual one. 

But, on the other hand, it confuses formal constituency (as set out in IFG2 Section 2.1 and IFG3 Section 1.1.3) with function structures. To explain, formal constituency is modelled as the grammatical rank scale, such that clauses consist of groups and phrases, which consist of words, which consist of morphemes. In contrast, the function structure of a rank unit, such as clause, is the relation between the structural elements, each of which is the function of a clause constituent, group or phrase (this being more precisely the case for experiential structures than the others, as suggested above).

[2] Trivially, contrary to Table 5.4, nominal and verbal groups have multivariate as well as univariate structure.

Seriously Misrepresenting The Logical Structure Of The Verbal Group

Martin, Matthiessen & Painter (2010: 164):

In addition, this Head may be expanded by other words; these serve the general function of Modifier. In the most expanded examples above, there are several instances of the Modifier function:

… To show this step-wise modification, we can represent the structure as a series of linked elements denoted by small Greek letters, with α for the Head and β, γ, δ etc. for successive Modifiers. For example:

Blogger Comments:

To be clear, the representation of the logical structure of the nominal group is potentially misleading. Halliday (1994: 191) represents the structure as:

On the other hand, the representation of the logical structure of the verbal group is seriously misleading in what it portrays as elements and by including the word functioning as Event in the recursive relation. Cf Halliday (1994: 200):

To be clear, the logical structure of the verbal group realises the system of tense. Because of this, it is realisation of the primary tense that functions as Head and the realisation of secondary tenses that function as Modifier. 

Halliday (1994: 199, 198) clarifies the realisations of tense:

and illustrates such realisations of tense as structural elements:

And it is because the logical structure realises tense as a recursive relation that the Event is not an element in the notation. Halliday & Matthiessen (2014: 411n):

A major point of difference between the verbal group and the nominal group is that the Event (unlike the Thing) is not the point of departure for the recursive modifying relationship. Hence it does not figure as an element in the notation.

The experiential and logical structures of this verbal group are actually as follows:

Misrepresenting A Verbal Group Complex As A Verbal Group

Martin, Matthiessen & Painter (2010: 162, 163):

 …

Blogger Comments:

This is misleading because it twice misrepresents the verbal group complex has decided to create as a verbal group.

Misrepresenting The Interpersonal Range Of The Medium

Martin, Matthiessen & Painter (2010: 158):

Thus we can now revise the table of correspondences between transitivity functions and interpersonal functions. See Table 4.23. The interesting result is that the participant function we have called the Medium stands out among the participant functions: it is the only one that serves as Subject or Complement but not as Adjunct. Agent, Beneficiary and Range are closer to circumstances in that they can also be mapped onto Adjunct. 

Blogger Comments:

This is misleading because it misrepresents the interpersonal variability of the Medium. Medium conflates with Adjunct in medio-receptive and locative-receptive clauses. Halliday & Matthiessen (2014: 342, 345, 351):