Metaphorical Realisations Of Speech Functions

Halliday & Matthiessen (2014: 700-1):

The command ‘vote against ...’ is realised metaphorically by a hypotactic clause nexus; it is realised as if it was a report of what the speaker says. This is just like the metaphorical realisation of modality of the explicitly subjective orientation. Thus the reported command can be tagged: I urge you to vote against ... will you? In other words, just like modality, speech function can be represented as a substantive proposition in its own right; and this proposition is a figure of sensing or saying that projects the original [i] proposal or [ii] proposition.

Interpersonal Projection: Modality

Halliday & Matthiessen (2014: 700-1):

But the notion of interpersonal projection is not, in fact, limited to modal assessment. Consider the short persuasive text in Text 10-5:

Text 10-5: Exploring – arguing: open letter appealing to members of an organisation [Text 6]

I don’t believe that endorsing the Nuclear Freeze initiative is the right step for California Common Cause. Tempting as it may be, we shouldn’t embrace every popular issue that comes along. When we do so we use precious limited resources where other players with superior resources are already doing an adequate job. Rather, I think we will be stronger and more effective if we stick to those issues of governmental structure and process, broadly defined, that have formed the core of our agenda for years. Open government, campaign finance reform, and fighting the special interests and big money – these are our kinds of issues. 

Let’s be clear: I personally favour the initiative and ardently support disarmament negotiations to reduce the risk of war. But I don’t think endorsing a specific freeze proposal is appropriate for CCC. We should limit our involvement in defence and weaponry to matters of process, such as exposing the weapons industry’s influence on the political process. Therefore, I urge you to vote against a CCC endorsement of the nuclear freeze initiative.

Expressions of modulation are bolded and expressions of modalisation are italicised; metaphorical expressions are underlined. Subjective assessment of modality permeates the text, like a prosody. In the case of modalisation, it is largely explicit; in the case of modulation, it is implicit until the very last clause. This clause realises the nucleus of the whole text; it is here that the key proposal is presented with a nuclear punch… . 

But what is the key proposal? One variant of I urge you to vote ... is you must vote ..., the two being relatable as explicit and implicit variants of subjective, high modulations of the type obligation. However, we can take one step further in the analysis. What the author is saying is Vote against ..., which is an ‘imperative’ clause – the congruent realisation of a proposal of the subtype ‘command’. This indicates the connection between ‘imperative’ clauses and modulation, which is why modulation was characterised as the ‘imperative type’ of modality. On the one hand, an ‘imperative’ clause imposes an obligation; on the other hand, the imperative tag checks the addressee’s inclination to comply (will you?). But the example also illustrates the connection between mood and projection.

Interpersonal Projection

Halliday & Matthiessen (2014: 700):

There is thus a fundamental relationship between modal assessment, including modality, and projection. To bring this out, we can interpret modal assessment as interpersonal projection. Interpersonal projection always involves the speaker or addressee as ‘projector’: ‘I think’, ‘I say’; ‘do you think’, ‘do you say’. It is always implicit unless it is made explicit through grammatical metaphor, by ‘co-opting’ ideational resources to do interpersonal service.

Metaphor Opens Up New Systemic Domains Of Meaning

Halliday & Matthiessen (2014: 699-700):

This is the general effect of grammatical metaphor: it construes additional layers of meaning and wording. To capture this layering in our grammatical analysis, we introduce one or more additional structural layers in the box diagram. The representation of grammatical metaphor in such diagrams shows how the metaphor is embodied in the structural organisation as an increase in the layers of meaning and wording. 

But there is, of course, also a systemic effect. Systemically, metaphor leads to an expansion of the meaning potential: by creating new patterns of structural realisation, it opens up new systemic domains of meaning. And it is the pressure to expand the meaning potential that in fact lies behind the development of metaphorical modes of meaning. Thus in the system of MODALITY, the system of ORIENTATION is expanded by the addition of a systemic contrast in MANIFESTATION between ‘explicit’ and ‘implicit’: the metaphorical modalities described above make it possible to make the orientation explicit in wordings such as I think and it is likely that, which, in turn, makes it possible to increase the delicacy of differentiations (cf. I think/imagine/expect/assume/suppose/reckon/guess; I would think/I would have thought; I imagine/I can imagine; and so on). 

As we have already seen, this same principle extends beyond modality and applies to modal assessment more generally (e.g. I regret, it is regrettable that). The metaphoric strategy is to upgrade the interpersonal assessment from group rank to clause rank – from an adverbial group or prepositional phrase serving within a simple clause to a clause serving within a clause nexus of projection.

The Extension Of Modality Through Metaphor

Halliday & Matthiessen (2014: 698-9):

As we have seen, the semantic domain of modality is extended through grammatical metaphor to include explicit indications of subjective and objective orientation: a modal proposition or proposal is realised, as if it was a projection sequence, by a nexus of two clauses, or as if it was a fact embedded as a Carrier in a relational clause with a modal Attribute, rather than by a single clause. Here the modal assessment itself is given the status of a proposition in its own right; but because the projecting clause of the nexus is metaphorical in nature, standing for an interpersonal assessment of modality, it is also, at the same time, a modal Adjunct in the clause realising the proposition/proposal.

Example Analysis Of Modal And Transitivity Metaphors

Halliday & Matthiessen (2014: 698, 699):

Figure 10-9 gives a further example, containing both an interpersonal metaphor and one of an ideational kind.

 

The Apparent Paradox On Which The Modality System Rests

Halliday & Matthiessen (2014: 698):

The importance of modal features in the grammar of interpersonal exchanges lies in an apparent paradox on which the entire system rests – the fact that we only say we are certain when we are not. If unconsciously I consider it certain that Mary has left, I say, simply, Mary’s left. If I add a high value probability, of whatever orientation, such as Mary’s certainly left, I’m certain Mary’s left, Mary must have left, this means that I am admitting an element of doubt – which I may then try to conceal by objectifying the expression of certainty. Hence whereas the subjective metaphors, which state clearly ‘this is how I see it’, take on all values (I’m sure, I think, I don’t believe, I doubt, etc.), most of the objectifying metaphors express a ‘high’ value probability or obligation – that is, they are different ways of claiming objective certainty or necessity for something that is in fact a matter of opinion. Most of the ‘games people play’ in the daily round of interpersonal skirmishing involve metaphors of this objectifying kind.

Metaphorical Realisations Of Modality

Halliday & Matthiessen (2014: 698):

The explicitly subjective and explicitly objective forms of modality are all strictly speaking metaphorical, since all of them represent the modality as being the substantive proposition. Modality represents the speaker’s angle, either on the validity of the assertion or on the rights and wrongs of the proposal; in its congruent form, it is an adjunct to a proposition rather than a proposition in its own right. Speakers being what we are, however, we like to give prominence to our own point of view; and the most effective way of doing that is to dress it up as if it was this that constituted the assertion (‘explicit’ I think ... ) – with the further possibility of making it appear as if it was not our point of view at all (‘explicit objective’ it’s likely that ... ).

Subjective vs Objective Orientation

Halliday & Matthiessen (2014: 696, 698):

The general difference in meaning between the subjective and the objective orientation can be seen from the effect of the tag. Compare the following two clauses:

he couldn’t have meant that, could he?
surely he didn’t mean that, did he?

In the first, the speaker wants the listener to confirm his estimate of the probabilities: ‘I think it unlikely; do you share my opinion?’. In the second, he wants the listener to provide the answer: ‘I think it unlikely, but is it in fact the case?’. It is possible to switch from a subjectively modalised clause to a non-modalised tag, as in this exchange in a store selling children’s books:

What do you reckon would be good for a five-year-old kid?
– She’ll like fairy tales, does she?

Here the salesperson’s reply means ‘I think it likely she likes fairy tales; is that the case?’ – whereas she’ll like fairy tales, will she? would have meant ‘do you agree that it is likely?’. The speaker is assuming, in other words, that the customer knows the preferences of the child; there would be no point in simply exchanging opinions on the subject.

Readiness: Ability vs Inclination

Halliday & Matthiessen (2014: 696):

There is one further category that needs to be taken into account, that of ability/potentiality, as in she can keep the whole audience enthralled. This is on the fringe of the modality system. It has the different orientations of subjective (implicit only) realised by can/can’t, objective implicit by be able to, and objective explicit by it is possible (for ... ) to. In the last of these, the typical meaning is ‘potentiality’, as in it was possible for a layer of ice to form. In the subjective it is closer to inclination; we could recognise a general category of ‘readiness’, having ‘inclination’ and ‘ability’ as subcategories at one end of the scale (can/is able to as ‘low’-value variants of will/is willing to). In any case can in this sense is untypical of the modal operators: it is the only case where the oblique form functions as a simple past, as in I couldn’t read that before; now with my new glasses I can.

Key To 30 Of The 144 Categories Of Modality

Halliday & Matthiessen (2014: 695-6):

30 Of The 144 Categories Of Modality Exemplified

Halliday & Matthiessen (2014: 695, 697):

This generates a set of 4 × 4 × 3 × 3 = 144 categories of modality. Thirty of these are illustrated in Figure 10-8.

Outer Value Modality

Halliday & Matthiessen (2014: 693-4):

With the outer values, on the other hand, if the negative is transferred the value switches (either from high to low, or from low to high):

Median Value Modality

Halliday & Matthiessen (2014: 693):

The median value is clearly set apart from the two ‘outer’ values by the system of polarity: the median is that in which the negative is freely transferable between the proposition and the modality:

Three ‘Values’ Of Modality

Halliday & Matthiessen (2014: 693-4):

The third variable in modality is the VALUE that is attached to the modal judgment: high, median or low. These values are summarised in Table 10-9, with ‘objective implicit’ forms as category labels.

Modality Examples By Type Of Orientation

Halliday & Matthiessen (2014: 692-3):

Examples of the combination of orientation and type are given in Table 10-8.

 

System Of Types Of Orientation In Modality

Halliday & Matthiessen (2014: 692):

The basic distinction that determines how each type of modality will be realised is the ORIENTATION: that is, the distinction between subjective and objective modality, and between the explicit and implicit variants… . The system is as in Figure 10-7.

 

These combine with all four types of modality, but with gaps; for example, there are no systematic forms for making the subjective orientation explicit in the case of usuality or inclination (i.e. no coded expressions for ‘I recognise it as usual that ... ’ or ‘I undertake for ... to ... ’). This is a systematic gap; these particular combinations would represent semantic domains where the speaker cannot readily pose as an authority.

Examples Of The Four Types Of Modality

Halliday & Matthiessen (2014: 692):

Here is an example of each of the four types:

1.i [probability]      There can’t be many candlestick-makers left.
1.ii [usuality]          It’ll change right there in front of your eyes.
2.i [obligation]       The roads should pay for themselves, like the railways.
2.ii [inclination]     Voters won’t pay taxes any more.

As these examples show, the modal operators can occur in all four types. Their use is more restricted in usuality and in inclination than in the other two types; but as a class they cover all these senses. This brings out what it is that the four types of modality have in common: they are all varying degrees of polarity, different ways of construing the semantic space between the positive and negative poles.

All Modalities Are Realised As Indicative Mood

Halliday & Matthiessen (2014: 692):

Note that modulation refers to the semantic category of proposals; but all modalities are realised as indicative (that is, as if they were propositions). Thus imperative go home!, when modulated, becomes indicative you must go home! (In philosophical semantics and linguistic accounts influenced by ‘modal logic’, probability is referred to as ‘epistemic’ modality (from Greek episteme, for ‘knowledge’) and obligation as ‘deontic’ modality (from Greek deon, for ‘what is binding’). Readiness has been referred to as ‘dynamic’ modality, and deontic and dynamic modalities have been grouped together as ‘root’ modality – our modulation type of modality; but usuality tends to be left out of accounts of modality.

The Four Types Of Modality

Halliday & Matthiessen (2014: 691):

The four types are set out in diagrammatic form in Figure 10-6. 

The System Of Modality

Halliday & Matthiessen (2014: 691):

The reason this area of the semantic system is so highly elaborated metaphorically is to be found in the nature of modality itself. … Modality refers to the area of meaning that lies between yes and no – the intermediate ground between positive and negative polarity. What this implies more specifically will depend on the underlying speech function of the clause. (1) If the clause is an ‘information’ clause (a proposition, congruently realised as indicative), this means either (i) ‘either yes or no’, i.e. ‘maybe’; or (ii) ‘both yes and no’, i.e. ‘sometimes’; in other words, some degree of probability or of usuality. (2) If the clause is a ‘goods-&-services’ clause (a proposal, which has no real congruent form in the grammar, but by default we can characterise it as imperative), it means either (i) ‘is wanted to’, related to a command, or (ii) ‘wants to’, related to an offer; in other words, some degree of obligation or of inclination. We refer to type (1) as MODALISATION and to type (2) as MODULATION; this gives a system as in Figure 10-5.

Expressions Of Probability In The Clause

Halliday & Matthiessen (2014: 688-9):

There is, in fact, a wide range of variants for the expression of modality in the clause, and some of these take the form of a clause nexus. This range is comparable to what we found with other types of modal assessment. If we limit ourselves first to the meaning of ‘probability’, the principal categories are as shown in Table 10-7; …

What happens is that, in order to state explicitly that the probability is subjective, or alternatively, at the other end, to claim explicitly that the probability is objective, the speaker construes the proposition as a projection and encodes the subjectivity (I think), or the objectivity (it is likely), in a projecting clause. (There are other forms intermediate between the explicit and implicit: subjective in my opinion, objective in all probability, where the modality is expressed as a prepositional phrase, which is a kind of halfway house between clausal and non-clausal status.)

An Interpersonal Metaphor Of Modality

Halliday And Matthiessen (2014: 686):

With modality, it is very clear that certain grammatical environments constitute metaphorical realisations of modality. … (see Figure 10-3). 

Here the cognitive mental clause I don’t believe is a metaphorical realisation of probability: the probability is realised by a mental clause as if it was a figure of sensing. Being metaphorical, the clause serves not only as the projecting part of a clause nexus of projection, but also as a mood Adjunct, just as probably does. The reason for regarding this as a metaphorical variant is that the proposition is not, in fact, ‘I think’; the proposition is ‘it is so’. This is shown clearly by the tag; if we tag the clause I think it’s going to rain we get I think it’s going to rain, isn’t it? not I think it’s going to rain, don’t I?. In other words the clause is a variant of it’s probably going to rain (isn’t it?) and not a first-person equivalent of John thinks it’s going to rain, which does represent the proposition ‘John thinks’ (tag doesn’t he?). …

What’s happened here is that there has been a realignment in the realisational relationship between semantics and grammar. … in examples such as the one analysed in Figure 10-3, a modalised proposition is realised as if it was a sequence, by a clause nexus of projection. The effect is that the modality and the modalised proposition are separated, each being realised by a clause in its own right: the modality is realised by the projecting mental clause and the proposition by the projected idea clause.

Manifestations Of Interpersonal Assessment: Clause vs Group

Halliday And Matthiessen (2014: 686):

The range of assessments assigned to propositions within the domain of the clause and the range of assessments assigned to things within the domain of the nominal group are not, of course, the same. They overlap; but there are kinds of assessment specific to the realm of propositions, just as there are other kinds specific to the realm of things. The common foundation is that they are both projections of the speaker’s assessment. This explains why clausal assessment can in fact be transformed into nominal assessment, as in disappointingly they forecast that ... => the disappointing forecast.

Modal Assessment: Domains Of Realisation

Halliday And Matthiessen (2014: 679-85):

The full range of modal assessment is shown in Table 10-6. The table indicates the different realisational domains of modal assessment, and it gives an indication of where categories of modal assessment correspond to categories of appraisal as described by Peter White and others. We have also noted assessments that may be realised as the Epithet of a nominal group. Such nominal groups serve as the Attribute of an intensive attributive relational clause with a fact as Carrier – the explicitly objective form of assessment. However, such Epithets of assessment may be assigned to things as well as to metathings: either the nominal group in which the Epithet serves is the Attribute assigned to a Carrier denoting a thing rather than a metathing, or else the Epithet is assigned directly to the Thing of the nominal group in which it serves.

Projection And Interpersonal Assessment

Halliday And Matthiessen (2014: 679):

Projection is thus dispersed across different grammatical environments. But the different manifestations of projection are not, of course, synonymous. The ideational manifestations make explicit the orientation of the assessment: the logical manifestation is explicitly subjective (e.g. I regret) whereas the experiential manifestation is explicitly objective (e.g. it’s regrettable). In contrast, the interpersonal manifestation leaves the orientation implicit:

ideational, logical – nexus:      explicit orientation: subjective (I regret)
ideational, exper. – 2 clauses: explicit orientation: objective (it’s regrettable)
interpersonal – 1 clause:          implicit orientation (regrettably)

The difference between ‘subjective’ and ‘objective’ orientation in the ideational manifestation follows from the general difference between a projecting ‘mental’ or ‘verbal’ clause with a Senser or Sayer and a ‘relational’ clause without such a ‘projector’. When the assessment is explicitly ‘subjective’, the Senser or Sayer has to be the speaker I if the clause is ‘declarative’ or the addressee you if it is ‘interrogative’ (as in Do you regret that it’s taken so long for you and your dad to get to work together?). If it is a person other than the speaker or addressee, the clause will still be a projecting one; but it will not be agnate with an interpersonal assessment.

Interpersonal Manifestations Of Projection: Subjective vs Objective Modal Assessment

Halliday And Matthiessen (2014: 678):

… projection includes both hypotactic projection of ideas/reports and pre-projected facts serving in a ‘mental’ or ‘relational’ clause. A hypotactic projection is always ‘subjective’; the speaker is represented explicitly as the Senser (e.g. I presume) or Sayer (e.g. I admit). A pre-projected fact in a ‘mental’ clause is like hypotactic projection in representing the assessment as ‘subjective’ – the speaker is explicitly represented as the Senser (e.g. I regret). In contrast, a pre-projected fact in a ‘relational’ clause represents the assessment as ‘objective’ (e.g. it is regrettable); but it may be ‘prefaced’ by a ‘mental’ clause with a ‘subjective’ orientation [e.g. I think it is regrettable].

Ideational vs Interpersonal Manifestations Of Projection

Halliday And Matthiessen (2014: 676-8):

When the need arises in discourse to attribute information to some source, this can be done logically by means of clause nexus of projection; but it can also be done experientially by means of a circumstance of Angle. …

Projection can also be manifested interpersonally in the form of a modal Adjunct … Unlike the logical and experiential manifestations, the interpersonal manifestation does not represent the Sayer or Senser; rather it enacts the speaker’s opinion – an enactment of his or her degree of commitment to the proposition: the proposition is assessed as being projected by somebody other than the speaker. This type of assessment is known as ‘evidentiality’: the modal Adjunct is used to indicate the evidential status of the proposition. The nearest logical equivalent would be ‘people say/they say that ...’, or ‘I hear that’: evidentiality is related to ‘verbal’ clauses and ‘mental’ clauses of perception.

In Chapter 4, we called this kind of modal assessment presumption; the comment Adjunct is realised by adverbial groups with adverbs as Head such as evidently, supposedly, reportedly, allegedly; arguably; presumably and they may have cognate verbs such as report, allege, argue serving as Process in a ‘verbal’ clause or suppose, presume serving as Process in a ‘mental’ clause. …

Projection is thus manifested interpersonally as modal assessment of the presumption type; but it extends beyond presumption to cover quite a few other types of modal assessment as well.

Grammatical Manifestations Of Projection vs Expansion

Halliday And Matthiessen (2014: 676):

Like expansion, projection is manifested both logically and experientially within the ideational metafunction; but outside the ideational domain, it is manifested interpersonally rather than textually, thus contrasting with the textual manifestation of expansion: see Table 10-5.

That is, while there are conjunctions marking rhetorical relations of elaboration, extension and enhancement, there are no conjunctions marking relations of quoting or reporting; and while there are interpersonal resources for realising projection (e.g. allegedly ‘they allege’; probably ‘I think’), there are no interpersonal Adjuncts or other interpersonal manifestations of expansion. There is thus a significant difference between expansion and projection.

Interpersonal Differences In Grammatical Manifestations Of Cause

Halliday And Matthiessen (2014: 675-6):

Just as the different domains of manifestation are textually distinct, they are also interpersonally distinct. When the domain of manifestation is a cohesive sequence of clauses, or a paratactic nexus of (free) clauses, the two figures related by expansion are enacted interpersonally as propositions or proposals. This means that each of them can be negotiated in their own right – accepted or denied, complied with or refused, and so on, as in she didn’t know the rules – oh yes, she certainly did. The same is true of the dominant (α) clause of a hypotactic nexus, since if it is a free clause, it realises a negotiable proposition or proposal. However, while the dependent (β) clause supports a proposition or proposal, it does not constitute one itself ; and if it is non-finite, it is even further removed from the realm of negotiation. A causal dependent clause (e.g. because she didn’t know the rules) is thus not presented as directly accessible to negotiation; it has to be accepted without argument.

When the domain of manifestation is a simple clause, there is just a single proposition rather than two. This obviously restricts the scope for negotiation; but when the causal relation is construed within the Process, it has become propositionalised or proposalised, and can be assessed … and negotiated … Here it is no longer the cause or the effect that is held up for negotiation but rather the causal relation. Being construed as nominal groups, the cause and the effect are not negotiable at all.

Textual Differences In Grammatical Manifestations Of Cause

Halliday And Matthiessen (2014: 674-5):

The different domains of manifestation of expansion, illustrated for cause in Table 10-4, are textually distinct. When the domain of manifestation is a cohesive sequence of clauses or a nexus of tactically related clauses, the conjunction group realising the causal relation is given a textual status within the clause in which it appears: it is either textual Theme or part of the Rheme. While a conjunction group with a structural conjunction such as so or because as Head is obligatorily thematic, there is a choice for conjunction groups with a cohesive conjunction such as consequently as Head …

In a hypotactic nexus, there is a further textual contrast that is not open to cohesive sequences and paratactic nexuses: the dependent β clause representing the cause may be either thematic or rhematic within the clause nexus; …

When the domain of manifestation of expansion is a simple clause, the potential textual status of the manifestation of the cause depends on how it manifested – (1) as minor Process within a prepositional phrase serving as a circumstance of Cause, (2) as Process, or (3) as Thing within a nominal group serving as a participant in a circumstantial relational clause.

(1) When it is manifested within a circumstance of Cause, the cause may be given the status of either Theme or Rheme, and, if it is Theme, it may be given the status of predicated Theme; …

(2) When the causal relation is manifested as Process (either in a hypotactic verbal group complex in a clause of any process type, or as the nuclear process in a circumstantial relational clause), its textual status will most likely be rhematic. More specifically, it is likely to be (part of) the transition between Theme and New. …

(3) When the cause is manifested as the Thing in a nominal group serving as a participant, it will have the thematic status assigned to that nominal group as a whole – either thematic or rhematic. But, in addition, it will be within the domain of operation of another textual system – the system of REFERENCE. This means that it is given a textual status as a discourse referent – either recoverable (identifiable) or non-recoverable (non-identifiable), and that it can be tracked in the development of the discourse.

Ideational Differences In Grammatical Manifestations Of Cause

Halliday And Matthiessen (2014: 673-4):

From an ideational point of view, the difference in meaning relates most directly to the question of what is construed as a quantum of change in the flow of events. The examples in Table 10-4 form a scale extended between two poles. At one pole, the experience of the flow of events is construed as two distinct quanta of change, realised by two independent clauses that are related cohesively but not structurally (She didn’t know the rules. Consequently, she died.). At the other pole, the experience of the flow of events is construed as a component part of a quantum of change – a participant that can itself be an element of some other quantum of change (cf. the quite incredible ignorance on which it is based must be a cause for grave concern), realised by a single nominal group (her death through ignorance of the rules). Intermediate between these two poles are various manifestations that represent a move from two distinct quanta of change via two interdependent ones to a single one.

The scale is thus one of degree of integration of two quanta of change. This scale of integration is based on the rank scale. At one pole, the sequence of cohesively related clauses transcends the rank scale, and, at the other, the nominal group is located at the rank below that of the clause. These two poles are thus connected by a move down the rank scale. At the same time, this move involves a shift in metafunction: textual – logical – logical + experiential – experiential. Here the meaning of expansion changes with the change of metafunctional manifestation. For example, the manifestation of cause changes from rhetorical relation (textual: consequently) via logico-semantic relation (logical: so, because) to process or minor process and even participant (experiential: cause, through; cause). This means that the category meaning of ‘cause’ changes; so while, for example, consequently and through share the meaning of cause, they differ in the category meanings they assign to it. (We shall see below that some of these realisations of cause are, in fact, ideational metaphors within the domain of the grammar.)

Grammatical Manifestations Of Cause

Halliday And Matthiessen (2014: 669, 673):

… for any given type of expansion we want to express, we have at our disposal a range of resources. For example, if we want to express an enhancing relationship of cause, the opportunities made available by the grammar include the one set out in Table 10-4.

These patterns of wording are agnate, but just like agnate patterns in general they are not synonymous: agnation always embodies both similarity and difference. The similarity is the basis for interpreting the patterns as alike, bringing them together within a paradigm, while the difference is the basis for treating them as variant types rather than as tokens of the same type. The patterns are similar in that they are all manifestations of different kinds of expansion, as illustrated by the examples of the enhancing notion of cause above. But how do the agnate variants differ in meaning? As always with questions of meaning, the answer can be found in its metafunctional organisation: differences turn out to be (i) ideational, (ii) textual and (iii) interpersonal.

Synoptic Summary Of Expansion

Halliday And Matthiessen (2014: 669-72):

Let us now present a systematic and comprehensive summary of the different grammatical environments in which elaboration, extension and enhancement are manifested: see Table 10-3. As the table shows, the environments of manifestation can be differentiated in terms of (i) metafunction – textual (CONJUNCTION), logical (INTERDEPENDENCY; MODIFICATION) and experiential (CIRCUMSTANTIATION; PROCESS TYPE: relational), and (ii) rank – clause and group/phrase. (The table could, in fact, be extended downwards along the rank scale to take account of patterns below the rank of group/phrase within the logical metafunction: word and morpheme complexes also embody interdependency relations that combine with expansion.)

From a grammatical point of view, the environments set out in Table 10-3 are, of course, all different. But seen from above, from the vantage point of semantics, they are all agnate ways of construing expansion. Collectively they thus construe expansion as a semantic system.

Logical vs Experiential Manifestations Of Expansion

Halliday And Matthiessen (2014: 666-7):

… the three subtypes of expansion (elaboration, extension and enhancement) combine with tactic relations to link one clause to another in the formation of clause complexes. The same pattern was found to operate in the formation of group and phrase complexes. In clause complexes and in group/phrase complexes, expansion is manifested within the logical mode of the ideational metafunction.

But we had already met two experiential manifestations of expansion within the system of transitivity. On the one hand, expansion is manifested in the augmentation of the clause by circumstances: these circumstantial augmentations cover all three types of expansion, with enhancement being the most highly developed one. On the other hand, expansion is manifested in the process of a ‘relational’ clause: ‘intensive’ clauses embody elaboration, ‘possessive’ clauses extension, and ‘circumstantial’ clauses enhancement. In the domain of the nominal group, there are also various manifestations of the three types of expansion. Thus the Qualifier may elaborate, extend or enhance the Thing;

Dispersed vs Compact Grammatical Realisations Of Semantic Systems

Halliday & Matthiessen (2014: 666, 667):

The semantic system of EXPANSION may be contrasted in this respect with that of FIGURATION: see Figure 10-2. The grammatical realisation of FIGURATION is ‘compact’, being confined to the TRANSITIVITY system of the clause. In contrast, the grammatical realisation of EXPANSION is ‘dispersed’, ranging over more than one grammatical unit. (We shall see later that compactly realised systems such as configuration may become dispersed in their realisation through the process of grammatical metaphor.)

 

Semantic Systems Realised In More Than One Grammatical Domain: Expansion And Projection

Halliday & Matthiessen (2014: 666):

There are two fundamental semantic domains of this kind – expansion and projection. For example, the expanding relation of ‘addition’ may be realised (1) cohesively by a conjunction such as also or (2) structurally by (a) an additive paratactic clause nexus marked by the structural conjunction and, (b) a circumstance of accompaniment marked by the preposition with or (c) an additive paratactic group nexus marked by and:

(1) She went to the market. Her son also went to the market.
(2a) She went to the market and so did her son.
(2b) She went to the market with her son.
(2c) She and her son went to the market.

These realisational variants are dispersed in the grammar, since they constitute different grammatical environments; but they are semantically agnate in that they all have the feature of ‘addition’. The fact that they are semantically agnate does not mean that they are synonymous; they share the feature ‘addition’ but they differ in other respects. For example, (2b) and (2c) construe one event, whereas (2a) and (1) imply two events. One semantic system, the system of EXPANSION, has thus evolved to bring together patterns of wording within grammatically distinct units, thereby extending its overall meaning potential.

Semantic Systems Realised In More Than One Grammatical Domain: Modality

Halliday & Matthiessen (2014: 666):

Every lexicogrammatical system realises some semantic system. For example, we have seen that the grammatical system of MOOD [clause] realises the semantic system of SPEECH FUNCTION [proposition/proposal]. Here one lexicogrammatical domain, the clause, corresponds to one semantic domain, the proposition/proposal. In addition, there are semantic systems that are realised by grammatical systems operating in more than one place in the grammar. These semantic domains range over two or more grammatical domains, spanning two or more grammatical units, as in the case of MODALITY noted above.

Metaphorical Realisation

Halliday & Matthiessen (2014: 665):

(ii) On the other hand, there are realignments in the realisational relationship between semantic units and grammatical ones. According to Table 10-1, a sequence is realised by a clause complex; the combination of a figure, a proposition/proposal and a message is realised by a clause; and we could add further realisational correspondences: for example, a participant is realised by a nominal group, a process by a verbal group and a circumstance by an adverbial group or a prepositional phrase. But once these couplings between the two strata of the content plane have been established, ‘cross-couplings’ become theoretically possible. 

For example, while sequences are realised by clause complexes and figures by clauses, it is theoretically possible that, under certain conditions, sequences would be realised by clauses – that is, as if they were figures. This is the possibility of metaphorical realisation, which has been taken up in English and many other languages, creating a more complex relationship between semantics and lexicogrammar than the one shown in Table 10-1. 

For example, instead of saying we and this common homeland are spiritually and physically united, so we were deeply pained as we saw ..., we can say that spiritual and physical oneness we all share with this common homeland explains the depth of the pain we all carried in our hearts as we saw .... But the two forms of realisation are not, of course, synonymous, so the effect is one of expanding the meaning potential of the language.

Transgrammatical Semantic Domains

Halliday & Matthiessen (2014: 665):

(i) On the one hand, there are semantic domains that range over more than a single grammatical unit. Thus the semantic domain of modality is realised in more than one place in the grammar; for example, it is realised by ‘mental’ clauses such as I suppose and by ‘relational’ clauses such as it is possible, by verbal groups with finite modal operators such as may and by adverbial groups with modal adverbs such as perhaps. These modal patterns within different grammatical units are not interchangeable synonyms; they have distinct values within the overall semantic system of modality. As we shall see below, while the forms above can all realise ‘low probability’, I think is explicitly subjective, it is possible is explicitly objective, may is implicitly subjective and perhaps is implicitly objective. This means that the semantic system of modality is more extensive than the modal features of any one given grammatical unit would suggest; it is realised not by a single grammatical unit but by a range of units: semantic unit a ↘ grammatical units m, n & o.

Summary Of Semantic Units In Relation To Grammar

Halliday & Matthiessen (2014: 664, 665):

The account that we have outlined is summarised in Table 10-1. The table shows the intersection of two content strata (semantics and lexicogrammar) with the three metafunctions. Within each metafunction there are semantic patterns that are intermediate between (1) the local semantic units that are realised by the clause and the clause complex and (2) the global semantic unit, i.e. the text.

Table 10-1 shows how semantic units are mapped onto grammatical ones. The principle is that of rank-based constituency – semantic unit a ↘ grammatical unit m; the key grammatical unit is the clause, as shown diagrammatically in Figure 10-1. But while this is the foundation on which the relationship between semantics and lexicogrammar is based, there are two other principles affecting this relationship, making it more complicated but also extending the meaning potential of language: (i) transgrammatical semantic domains – domains of meaning extending across different grammatical units, and (ii) metaphor – incongruent realisational relations between semantics and lexicogrammar.

Intermediate Ideational Semantic Units: Episodic Patterns

Halliday & Matthiessen (2014: 663):

Ideational. Figures combine to form sequences; and these, in turn, may combine to form episodic patterns, as in narratives and other chronologically organised texts or chronological passages within other kinds of text.

Intermediate Interpersonal Semantic Units: Patterns Of Exchange

Halliday & Matthiessen (2014: 661, 663):

Interpersonal. Propositions/proposals combine to form patterns of exchange involving two or more interactants… . At least in some registers such as interviews, classroom discourse and quiz shows, the patterns are regular and predictable enough to appear to be compositional in nature: interpersonal units of exchange consisting of certain sequences of propositions/proposals. But whatever the status is of such exchange patterning, the critical point is that it is interpersonal rather than textual or ideational, and that it is intermediate between propositions/proposals and the whole text.

Intermediate Textual Semantic Units: Information Flow Patterns

Halliday & Matthiessen (2014: 661):

Textual. Messages combine to form periodic movements of information… . Such phases of messages help construct the flow of information as the text unfolds and we can refer to them simply as information flow patterns.

Intermediate Semantic Units

Halliday & Matthiessen (2014: 661):

What is intermediate in the semantics between text on the one hand and message, proposition/proposal and figure on the other hand? Messages, propositions/proposals and figures can combine with units of the same metafunctional type to form more extensive semantic patterns in the creation of text. These patterns are distinct for each metafunction.

Semantic Units Realised By The Clause

Halliday & Matthiessen (2014: 661, 662):

However, even though we have to leave open the question of whether there is a single semantic compositional scale or more than one, we can say something about the lower region of such a scale or such scales. This is the region where semantic units are realised by lexicogrammatical ones. We have already referred to these semantic units throughout the book, and the principle that emerges here is the familiar metafunctional one. The clause is a multifunctional construct in the grammar, one that realises three different semantic units, one for each metafunction: textual – message, interpersonal – proposition/proposal, and experiential – figure. These relationships are shown in Figure 10-1: the three semantic units deriving from the three metafunctions are all mapped onto the clause, which thus unifies the three metafunctional strands of meaning.

A Single Compositional Scale For Semantics?

Halliday & Matthiessen (2014: 660-1):

The upper bound of the semantic stratum is, as we have said, the text: this is the most extensive unit of meaning. The upper bound of the lexicogrammatical system is the clause: this is the most extensive unit of wording. In the grammar, there is a single, generalised compositional scale – the grammatical rank scale (clause – group/phrase – word – morpheme); and we can specify not only the upper and lower bounds of this scale – the clause and the morpheme, respectively – but also the intermediate units of patterning: group/phrase and word. But in the semantics, it is far from clear whether there is a single compositional scale: such a scale would have to be generalised across all registerial varieties of a language, but we know that texts vary considerably from one register to another. It is quite possible that different registers operate with different compositional scales; for example, one such scale was identified in the organisation of class room discourse by Sinclair & Coulthard (1975) and another in the organisation of certain types of conversation by Cloran (1994). This issue can only be settled after a great deal more research into the semantics of text has been carried out.

Text

Halliday & Matthiessen (2014: 660):

A text is thus a unit of meaning – more accurately, a unit in the flow of meaning that is always taking place at the instance pole of the cline of instantiation. How does this semantic unit relate to the units and unit complexes of grammar – the clause (clause complex), the groups (group complexes) and so on? The folk view is that a text consists of clauses (or ‘sentences’); but this is a misleading simplification of a more indirect – but much more flexible and powerful – relationship. A text does not ‘consist of’ clauses (clause complexes) – there is no part-whole or ‘constituency’ relationship between a text and a clause (complex) and there is no single rank scale with text and clause as ranks. Rather a text is realised by clauses (clause complexes), the two being located on different strata – semantics (the stratum of meaning) and lexicogrammar (the stratum of wording) respectively.

Blogger Comments:

It is important to understand (and maintain) the difference between a text as semantic unit, and text as an instance of language. A text as semantic unit is the highest unit on the semantic stratum, and so, realised by the clauses of lexicogrammar. A text as an instance of language is an instance of language as a whole, not just the semantic stratum.

Logogenetic Patterns And Cohesion

Halliday & Matthiessen (2014: 659-60):

Through the accumulation of logogenetic patterns and through the resources of cohesion, lexicogrammar makes a considerable contribution to the development of patterns in a text that extend beyond a single grammatical unit such as the clause, or even a complex of units such as the clause complex; and this is, of course, why lexicogrammatical analysis of a text can tell us so much about how it works. The patterns that are developed in this way are, however, patterns of meaning, not patterns of wording; they are patterns at the level of semantics rather than at the level of lexicogrammar. This is so because text is, as we have emphasised, a semantic phenomenon in the first instance; it is meaning unfolding in some particular context of situation. For example, the grammatical system of conjunction gives speakers and writers the resources to mark transitions in the development of a text – i.e., to mark rhetorical relations used to expand the text step by step; and the rhetorical relations that are marked in this way by conjunctions are semantic relations organising the text as a flow of meaning.

Blogger Comments:

A logogenetic pattern is a pattern of instantiation; that is, a pattern of feature selection and realisation statement activation. Different patterns identify different texts, and different patterns of patterns identify difference 'above' the instance pole of the cline of instantiation. A logogenetic pattern thus includes selections from the lexicogrammatical system of cohesion.

The System Of Cohesion

Halliday & Matthiessen (2014: 659):

… lexicogrammar has evolved textual resources for creating cohesive links that have the ability to indicate semantic relationships in the unfolding text beyond the domain of grammar. These resources are known collectively as the system of COHESION. Cohesion includes (1) the system of CONJUNCTION for marking textual transitions in the unfolding of text and (2) the systems of REFERENCE, ELLIPSIS & SUBSTITUTION, and LEXICAL COHESION for giving elements different textual statuses in the unfolding of text.

Logogenetic Patterns

Halliday & Matthiessen (2014: 659):

… in the course of unfolding of text, lexicogrammatical selections create logogenetic patterns at all ranks. This is patterning in the text that has nothing to do with composition or size: instead of composition (the relationship between a whole and its parts), the patterning is based on instantiation (the relationship between an instance and a generalised instance type). The patterning represents a slight move up this cline from the single instance to a pattern of instances, as in a news report where one projecting verbal clause after another is selected until this emerges as a favourite clause type. The logogenetic patterns that emerge as a text unfolds form a transient system that is specific to that text; but from repeated patterns over many such transient systems may, in turn, emerge a generalised system characteristic of a certain type of text or register.

The Complementarity Of Clause Complexing And Conjunction

Halliday & Matthiessen (2014: 655, 657):

These two resources complement one another in the grammatical realisation of transitions in text. The general principle of complementarity is this: clause complexing does relatively more work locally, while conjunction does relatively more work non-locally and even globally. Clause complexing ‘choreographs’ the local development of text by means of univariate structure, indicating both taxis and type of logico-semantic relation. Conjunction can work together with clause complexing, reinforcing local relations, but it tends to take over from clause complexing as the relations become less local and more global. Looked at from the point of view of lexicogrammar, this means that the local organisation tends to be ‘tighter’ whereas the more global organisation tends to be ‘looser’. At the same time, looked at from the point of view of context, the more global organisation is subject to more contextual guidance in the form of generic structure. In this way, grammar and context complement one another in their contributions to the semantic organisation of text. …

However, this pattern of complementarity varies significantly across registers. In particular, in unselfconscious spoken text, clause complexing does relatively more work and conjunction relatively less. Here clause complexes can extend well beyond local relations, reaching a span of up to 20 to 30 clauses. Conversely, in planned written text, clause complexing does relatively less work and conjunction relatively more.

Textual Transitions: Conjunction And Clause Complexing

Halliday & Matthiessen (2014: 655):

While textual statuses are assigned to components of messages, textual transitions hold between whole messages, or groups of whole messages. Such transitions may, of course, be left to the listener or reader to infer, without the help of any explicit markers; but they may also be marked explicitly by means of textual or logical resources. The textual resource for indicating such transitions is the system of CONJUNCTION. Here the type of relation is indicated, but there is no distinction between paratactic and hypotactic relations. The logical resource combines type of relation with TAXIS in the univariate structure of clause complexing.

Textual Status And Lexical Cohesion

Halliday & Matthiessen (2014: 654):

Unlike REFERENCE and ELLIPSIS, the resource of LEXICAL COHESION is, in general, neutral with respect to textual statuses; that is why we have not included it in Table 9-4. A given lexical item can be assigned any of the textual statuses; it is not predisposed to any particular status. Similarly, a given lexical relation is also neutral with respect to textual statuses. However, a particular lexical chain in a text is very likely to show some systematic pattern in relation to textual status.

Textual Statuses

Halliday & Matthiessen (2014: 652):

The systems of THEME, INFORMATION, REFERENCE and ELLIPSIS are all concerned with textual statuses. Speakers assign these statuses to components of messages to help themselves produce texts and to help their listeners interpret them. These textual statuses are independently variable. For example, thematic status may be combined with either given or new, and the same is true of rhematic status. However, there are certain unmarked combinations: in the unmarked case, Theme is Given and New falls within Rheme. The thematic statuses and the favoured combinations are shown in Figure 9-5.

Textual Grammatical Resources Viewed 'From Above'

Halliday & Matthiessen (2014: 650, 652):

Looked at ‘from below’, these textual resources fall into two categories – those that engender grammatical structure (theme and information) and those that do not (conjunction, reference, ellipsis, lexical cohesion). Looked at ‘from above’, these textual resources are concerned either with textual transitions between messages or with textual statuses of components (elements) of these messages. These two classificatory perspectives are intersected in Table 9-19. 

The table indicates that structural and cohesive resources work together in the marking of textual transitions and in the marking of textual statuses. In the latter case, all the resources are indeed textual; but in the former, the structural resources are logical rather than textual – the logical relations of clause complexing. … The table does not include lexical cohesion. This is like reference and ellipsis in that it involves components of messages rather than whole messages; but it often works together with conjunction in the creation of relations in text. We shall thus have occasion to refer to it both in our discussion of textual statuses and in our discussion of textual transitions.

The Textual Resources Of The Lexicogrammar Of English

Halliday & Matthiessen (2014: 650):

We have identified the following features as those which combine to make up the textual resources of the lexicogrammar of English:

(A) structural

1 thematic structure: Theme and Rheme
2 information structure and focus: Given and New

(B) cohesive

1 conjunction
2 reference
3 ellipsis (that is, ellipsis and substitution)
4 lexical cohesion

Collocation: Register Variation

Halliday & Matthiessen (2014: 650):

Notice, finally, that collocations are often fairly specifically associated with one or another particular register, or functional variety of the language. This is true, of course, of individual lexical items, many of which we regard as ‘technical’ because they appear exclusively, or almost exclusively, in one kind of text. But it is also noteworthy that perfectly ordinary lexical items often appear in different collocations according to the text variety. For example, hunting, in a story of the English aristocracy, will call up quarry and hounds (or, at another level, shooting and fishing); in an anthropological text, words like gathering, agricultural and pastoral; as well as, in other contexts, bargain, souvenir, fortune and such-like.

Collocation: Types Of Relation

Halliday & Matthiessen (2014: 649):

In general, the semantic basis of many instances of collocation is the relation of enhancement, as with dine + restaurant, table; fry + pan; bake + oven. These are circumstantial relationships (for collocations involving Process + Manner: degree, e.g. love + deeply, want + badly, understand + completely), but as the example with smoke + pipe illustrates, participant + process relationships also form the basis of collocation – the most important ones involving either Process + Range (e.g. play + musical instrument: piano, violin, etc.; grow + old) or Process + Medium (e.g. shell + peas, twinkle + star, polish + shoes); and there are also combinations involving functions in the nominal group, in particular, Epithet + Thing (e.g. strong + tea, heavy + traffic, powerful + argument) and Facet + Thing (e.g. pod + wales, flock + birds, school + fish, herd + cattle, gaggle + geese).

While we can typically find a semantic basis to collocation in this way, the relationship is at the same time a direct association between the words; if pipe is in the text then smoke may well be somewhere around, at least with considerably greater probability than if we just pulled words out of a hat on the basis of their overall frequency in the language. We get ready for it, so to speak; and hence if it does occur it is strongly cohesive (cf. Hoey’s, 2005, notion of lexical priming).

Collocation

Halliday & Matthiessen (2014: 648-9):

At the same time there are other instances of lexical cohesion that do not depend on any general semantic relationship of the types just discussed, but rather on a particular association between the items in question – a tendency to co-occur. This ‘co-occurrence tendency’ is known as collocation. For example,

A little fat man of Bombay
Was smoking one very hot day.
But a bird called a snipe
Flew away with his pipe,
Which vexed the fat man of Bombay.

There is a strong collocational bond between smoke and pipe, which makes the occurrence of pipe in line 4 cohesive. Clearly there is a semantic basis to a collocation of this kind; a pipe is something you smoke, and the words pipe and smoke are typically related as Range to Process in a behavioural process clause. Hence pipe here will be interpreted as ‘the pipe that he was smoking at the time’.

Blogger Comments:

To be clear, smoke is not a behavioural Process, if only because pipe is not a Behaviour (the Range of a behavioural Process).

Meronymy vs Hyponymy

Halliday & Matthiessen (2014: 648):

There is no very clear line between meronymy and hyponymy, especially with abstract terms; and a given set of items may be co-hyponyms of one term but co-meronyms of another – for example, chair, table, bed are ‘kinds’ (hyponyms) of furniture, but ‘parts’ (meronyms) of furnishings; forward, half-back, back are ‘kinds’ of players but ‘parts’ of a team, and so on. But since either relationship is a source of lexical cohesion it is not necessary to insist on deciding between them.

Meronymy: Examples

Halliday & Matthiessen (2014: 648):

The occurrence of any pair of items within the set will be cohesive; for example

||| On the left of the park lies the Exhibition Centre [[ which covers a massive 25,000 square metres of column-free space under the one roof]] . ||| Opened in January 1988, || the Centre is designed to hold major international exhibitions. ||| The glassed eastern facade is stepped back in five separate stages [[ that can be partitioned off to form smaller halls]]. ||| The fifth hall is linked by covered walkway to the Convention Centre. ||| 

||| Elfrida had a beautiful little glass scent-bottle. ||| She had used up all the scent long ago; || but she often used to take the little stopper out ||| ... 

||| She knelt down || and looked along the passage into the loveliest garden [[ you ever saw]] . ||| How she longed to get out of that dark hall, || and wander about among those beds of bright flowers and those cool fountains, || ... 

where roof, facade, hall are meronyms of centre, stopper is a meronym of bottle, and flowers and fountains are co-meronyms of garden.

Hyponymy vs Meronymy

Halliday & Matthiessen (2014: 647-8):

The general sense of hyponymy is ‘be a kind of’, as in ‘fruit is a kind of food’. There is a similar relation in the extending domain (Kinds of lexical relations playing a role in lexical cohesion). This is meronymy – ‘be a part of’. These two lexical relations are contrasted diagrammatically in Figure 9-4: given a lexical set consisting of either hyponyms, where x, y and z are all ‘kinds of’ a, or meronyms, where p, q and r are all ‘parts of’ b.

 

Hyponymy

Halliday & Matthiessen (2014: 646-7):

Repetition and synonymy are both elaborating relations based on identity; one lexical item restates another. There is a second kind of elaborating relationship – attribution. This is based on classification (specific to general): the first lexical item represents a class of thing and the second either (i) a superclass or a subclass or (ii) another class at the same level of classification. For example:

And do you know anything about mediæval literature; have you ever heard of any other kinds of literature in the mediæval period besides Chaucer? 

Most limestone probably originates from organisms that remove calcium carbonate from sea water. The remains of these animals may accumulate to form the limestone directly, or they may be broken and redeposited. 

Noah’s wife and his sons’ wives went to the fields to gather fruit and grain and vegetables. They would need plenty of food for themselves and the animals on the ark. The remains of these animals may accumulate to form the limestone directly, or they may be broken and redeposited.

Thus in the last example, fruit, grain and vegetables are co-hyponyms of food.

Antonymy

Halliday & Matthiessen (2014: 646):

A special case of synonymy is its opposite, antonymy. Lexical items which are opposite in meaning, namely antonyms, also function with cohesive effect in a text. For example, woke and asleep in

He fell asleep. What woke him was a loud crash.

Synonymy Without Identity Of Reference

Halliday & Matthiessen (2014: 646):

The occurrence of a synonym even where there is no particular referential relation is still cohesive; for example, see Text 9-11

Text 9-11: Recreating – narrating (written, monologic): limerick
There was a man of Thessaly
And he was wondrous wise.
He jumped into a hawthorn bush
And scratched out both his eyes.
And when he saw his eyes were out
With all his might and main
He jumped into a quickset hedge
And scratched them in again.

where the quickset hedge is not the same entity as the hawthorn bush but there is still cohesion between the synonyms hedge and bush.

Blogger Comments:

To be clear, referential relations are distinct from, and irrelevant to, the functioning of lexical cohesion.

Tracking A Participant Through The Discourse

Halliday & Matthiessen (2014: 645-6):

In fact we might have (disregarding the scansion, of course) any of the following sequences:

four-&-twenty blackbirds ... the blackbirds began to sing
                     "                     the birds began to sing
                     "                     the creatures began to sing
                     "                     they began to sing

the reference item they being simply the most general of all. Compare ankylosaur ... creature in the following description of a dinosaur:

As an added means of self-defence the ankylosaur had a club on its tail. The creature may have been able to swing the club with great force and aim a savage blow at an enemy.

Such instances are typically accompanied by the reference item the. This interaction between lexical cohesion and reference is the principal means for tracking a participant through the discourse. Instead of the creature we might simply have had the personal reference it; and earlier in the same text, we find such examples:

Ankylosaurus – the “fused lizard” – was the largest of the ankylosaurs, but in spite of its size and frightening appearance it fed only on plants.

Blogger Comments:

To be clear, the source of the notion of 'tracking a participant through the discourse' — which Matthiessen again fails to acknowledge — is not Halliday, but Martin (1992), further elaborated in Martin & Rose (2007). This misunderstanding derives from Martin's confusion of textual reference with ideational denotation (and deixis), and consequently, his confusion of reference with lexical cohesion. Evidence here (Martin 1992) and here (Martin & Rose 2007). In Martin's model, speakers are said to track themselves through a text, and participants are said to be tracked by being omitted.

Synonymy With Identity Of Reference

Halliday & Matthiessen (2014: 645-6):

Here the range of potentially cohesive items includes synonyms of the same or some higher level of generality: synonyms in the narrower sense, and superordinates. For example, in

Four-&-twenty blackbirds, baked in a pie.
When the pie was opened, the birds began to sing.

we have one instance of repetition (pie ... pie) and one of synonyms (blackbirds ... birds). birds, however, is at a higher level of generality than blackbirds; it is a superordinate term. …

Related to these are examples such as the following, where there is still identity of reference, although not to a participant, and the synonym may not be in the same word class (cheered ... applause; cried ... tears):

Everyone cheered. The leader acknowledged the applause.
I wish I hadn’t cried so much! I shall be punished for it, I suppose, by being drowned in my own tears!

Synonymy

Halliday & Matthiessen (2014: 645):

In the second place, lexical cohesion results from the choice of a lexical item that is in some sense synonymous with a preceding one; for example sound with noise, cavalry with horses in

He was just wondering which road to take when he was startled by a noise from behind him. It was the noise of trotting horses. ... He dismounted and led his horse as quickly as he could along the right-hand road. The sound of the cavalry grew rapidly nearer ...

Here again the cohesion need not depend on identity of reference. But once we depart from straightforward repetition, and take account of cohesion between related items, it is useful to distinguish whether the reference is identical or not, because slightly different patterns appear.

Lexical Item: Inflectional And Derivational Variants

Halliday & Matthiessen (2014: 644-5):

… in order for a lexical item to be recognised as repeated it need not be in the same morphological shape. For example, dine, dining, diner, dinner are all the same item, and an occurrence of any one constitutes a repetition of any of the others. Inflectional variants always belong together as one item; derivational variants usually do, when they are based on a living derivational process, although these are less predictable. (For example, rational and rationalise are probably still the same lexical item, though the relationship between them has become rather tenuous; but neither now goes with ration — rational is closer to reason, though not close enough to be considered the same item.)

In Landor’s line

I strove with none, for none was worth my strife

there is a strongly felt cohesion between strife and strove, suggesting that strive, strove and strife are one and the same lexical item.

Repetition

Halliday & Matthiessen (2014: 644):

The most direct form of lexical cohesion is the repetition of a lexical item; e.g. bear in

Algy met a bear. The bear was bulgy

Here the second occurrence of bear harks back to the first. In this instance, there is also the reference item the, signalling that the listener knows which bear is intended; and since there is nothing else to satisfy the the, we conclude that it is the same bear. But this referential link is not necessary to lexical cohesion; if we had Algy met a bear. Bears are bulgy, where bears means ‘all bears’, there would still be lexical cohesion of bears with bear. In this case, however, there would be only one tie; whereas in the example cited first there are two, one referential (the) and one lexical (bear).