Using (Misanalysed) Transitivity To Misunderstand (Rebranded) Ergativity

Martin & Rose (2007: 94):

Another two kinds of Range are a quality or a possession of the Medium.  In this case the process is one of ‘being' or ‘having’, that relates the quality or a possession to the Medium:

 

Blogger Comments:

[1] To be clear, here Martin & Rose confuse Halliday's particularised transitive model (attributive Processes) with Halliday's generalised ergative model, and use the former to elaborate the latter as their means of rebranding of Halliday's grammar as Martin's discourse semantics (nuclear relations).

[2] To be clear, 'quality' and 'possession' are the authors' rebrandings of Halliday's Attributes of intensive vs possessive clauses.

[3] To be clear, here Martin & Rose analyse the clause I was torn to pieces as attributive:

I	was	torn to pieces
Medium	Process	Range
Carrier	relational	Attribute

However, the Process, in this metaphorical clause, is was torn not was.  This is demonstrated by the agnate clauses:

• I was torn to pieces by the news.
• The news tore me to pieces.

On this basis, the metaphorical clause is material, with to pieces serving as Range (resultative Attribute):

I	was torn	to pieces	(by the news)
Medium	Process:	Range	Agent
Goal	material	resultative Attribute	Actor

Cf the material clause:

I	was shattered	(by the news)
Medium	Process:	Agent
Goal	material	Actor

as metaphorical for the mental clause:

I	was traumatised	(by the news)
Medium	Process:	Agent
Senser	mental	Phenomenon

Using The Transitive Model To Misunderstand The Ergative Model

Martin & Rose (2007: 94):

A process may also be instigated by the Medium and extended to a second participant that is not affected by the process, known as a Range. The first type of Range is an entity that the process extends to:

 

Blogger Comment:

[1] To be clear, this confuses Agent (the participant that is the external cause of the Process) with Medium (the participant through which the Process is actualised).  Moreover, Martin & Rose (p91) have previously defined the Agent in the same terms:

An Agent instigates the process, which affects the Medium in some way…

[2] To be clear, here Martin & Rose confuse two distinct theoretical meanings shared by the word 'extension'.  The first theoretical meaning is 'extension' as the basis of the transitivity model. Halliday & Matthiessen (2014: 226):

The concept of extension is in fact the one that is embodied in the classical terminology of ‘transitive’ and ‘intransitive’, from which the term ‘transitivity’ is derived. According to this theory the verb spring is said to be intransitive (‘not going through’) and the verb catch is said to be transitive (‘going through’ – that is, extending to some other entity).

That is, in this respect, Martin & Rose confuse the transitive model of the clause with the ergative model (their 'nuclear relations').

The second theoretical meaning is 'extension' as an expansion type.  The authors' claim — based on 'extension' as the transitivity principle — is that the relation of Range to Process is one of extension.  However, extension is the only logico-semantic relation that does not obtain between Range and Process.  Halliday & Matthiessen (1999: 168-9):

The Range is not some entity that is impacted by the Process; it either (i) expands the Process, or (ii) is projected by it. (i) Where the relationship is one of expansion, this take[s] one of two forms: the Range either (a) elaborates the Process in an objectified form, or (b) enhances it by delimiting its scope.

[3] To be clear, the relation obtaining between Range and Process in each of three instances is projection, not expansion: extension.  In the first, a Phenomenon (me) is related by projection to a mental Process (envied), while in the second and third, Verbiage (the pain and bitterness, the feared word) is related by projection to a verbal Process (can't explain, would mutter).

Nucleus		'	Range
Medium	Process
all my girlfriends	envied		me
I	can't explain	"	the pain and bitterness
they	would mutter		the feared word

The fact that these are projection relations is demonstrated by the agnate circumstance type in each case, which is Matter, a circumstance of projection:

• all my girlfriends were envious of me
• I can't explain about the pain and bitterness
• they would mutter about the feared word

Misconstruing Nuclearity As Predictability

Martin & Rose (2007: 93-4):

Lexical relations between the central elements of these activities are widely predictable across fields, i.e. the processes of exchanging (applying for, granting, giving, refusing). Relations between central and nuclear elements are predictable within the general field of granting amnesty (a common practice). But relations between these activities and more marginal elements are only predictable within the particular field of the Truth and Reconciliation Commission (e.g. who the Commission can and cannot grant amnesty to).

Blogger Comments:

[1] To be clear, what Martin & Rose term the 'central element', a Process, is an element of a grammatical structure, the clause, not of an "activity", and relations between such grammatical elements are logico-semantic, not lexical.

[2] Here Martin & Rose misunderstand the principle of nuclearity as one of predictability.  To be clear, in Halliday's model of clause ergativity, nuclearity measures the degree of participation in a process. Moreover, the proposed relative gradings of predictability are based on misunderstandings; see below.

[3] The claim here is that the relation between Processes like granting (central) and amnesty (nuclear) is predictable within the "general" field of granting amnesty.  There are four basic problems with this:

• Firstly, this is an instance of the logical fallacy known as circular reasoning: if granting + amnesty, then granting + amnesty.
• Secondly, the relation between central and nuclear elements is always the same, and thus always "predictable", irrespective of field, since it is a relation between clause elements (Process and Medium).
• Thirdly, Martin & Rose misunderstand the relation between Process and Medium as one of extension, in contradiction of the definition of Medium as the participant through which a Process is actualised.
• Fourthly, amnesty does not serve as Medium of the Process granting, 

[4] Trivially, the granting of amnesty is not a common practice; see for example the discussion here.

[5] The claim here is that the relation between Processes like granting and the participants Agent (e.g. the Commission) and Beneficiary (e.g. to those who plead guilty) is only predictable within the "particular" field of the Truth and Reconciliation Commission.  There are two basic problems with this:

• Firstly, the relation between central and marginal elements is always the same, and thus always "predictable", irrespective of field, since it is a relation between clause elements.
• Secondly, Martin & Rose mistake the Medium of the Process, the Commission, as Agent; i.e. as the cause of the granting process rather than the medium through which granting is actualised, and thus, as marginal ("only predictable within the particular field") rather than nuclear ("predictable").

[6] To be clear, here the meaning of activities has shifted again, this time from clause to Process.

[7] To be clear, on the authors' analysis, the Commission itself is just as marginal as those it can grant amnesty to.

Misunderstanding Clause Ergativity And Misapplying It To Data

Martin & Rose (2007: 93):

Within the field of granting amnesty, people participate in each activity, as Medium, Agents or Beneficiaries, i.e. as nuclear or marginal elements of the activity. To show their lexical relations we can use:

• symbols for nuclear relations: ‘=’ for central, ‘+’ for nuclear, and ‘x’ for marginal (following Halliday’s 1994/2004 symbols for logical relations),
• lexical rendering of pronouns and implicit elements,
• ‘=’ for relations between processes that are parts of a field, as follows:

Blogger Comments:

[1] Here again Martin & Rose confuse context (field) with language (clause functions).  In SFL theory, field is the ideational dimension of context (the instantiation cline from culture to situation).  In contrast, the ergative functions of Process, Medium, Agent and Beneficiary are features of language: semantic distinctions that are realised in the grammar of the clause.  That is, the authors confuse what's going on, culturally, in the creation of text, with what's going on in the text itself.

[2] To be clear, relations between elements of clause structure are grammatical relations, not lexical relations.

[3] To be clear, the claims here are that:

• the relation between the centre (Process) and itself is elaboration (=);
• the relation between the centre (Process) and the nucleus (Medium) is extension (+);
• the relation between the centre (Process) and the margin (Agent or Beneficiary) is enhancement (x).

Only the last of these three claims demonstrates an understanding of Halliday's model.  Of the first claim, it will be seen later here that Martin & Rose also include Range: process within the centre — though they also include intensive and possessive Attributes such as (be) an Englishman and (have) the guts, while, at the same time, excluding intensive and possessive Attributes such as (going) mad and (had) another lover.

Of the second claim, the notion that the Medium extends (adds to) the Process is inconsistent with the definition of the Medium as the participant through which the Process is actualised.  In Halliday's original model, the Process and Medium together form the experiential Nucleus of the clause.

[4] To be clear, relations between Processes of different clauses are neither lexical relations, nor nuclear relations within the clause. Moreover, in the examples provided, the relation between applying for and (not) giving is not one of elaboration, since the relation is not one of partial identity: i.e. one does not restate or clarify the other.

[5] Here again Martin & Rose misconstrue the realisation relation between two levels of symbolic abstraction, context (field) and language (processes), as one of composition ('parts of') at the same level.

[6] To be clear, the authors' schema of "lexical" relations not only misunderstands Halliday's ergative model but misapplies it to the data, misinterpreting the Medium (police officers, the Commission) as Agent, and misinterpreting the Range (amnesty) as Medium.  This can be made explicit by comparing the authors' analysis:

Margin	x	Centre	+	Nucleus	x	Margin
Agent		Process		Medium		Beneficiary
		granting		amnesty
police officers		apply for		amnesty
the Commission		gives		amnesty		to those who plead guilty
the Commission		gives		amnesty		to innocent people
the Commission		not give		amnesty		to those who claim to be innocent
the Commission		refused		amnesty		to the police officers who…

with an application of Halliday's ergative model:

Range	=	Nucleus		x	Beneficiary
		Medium	Process
amnesty		police officers	apply for
amnesty		the Commission	gives		to those who plead guilty
amnesty		the Commission	does not give		to innocent people
amnesty		the Commission	does not give		to those who claim to be innocent
amnesty		the Commission	refused		to the police officers who…

Confusing Context With Language

Martin & Rose (2007: 93):

We said earlier that a field consists of sequences of activities. The granting of amnesty is one activity within the Truth and Reconciliation field, that includes activities such as applying for, giving and refusing. This hierarchy of activities can be represented by a tree diagram, as in Figure 3.11. 

 

Blogger Comments:

[1] To be clear, even Martin & Rose model field as context — even if they misconstrue context as register — and here they locate activity sequences as discourse semantics, contradicting Martin (1992) where activity sequences are located in context.  That is, even in their own terms, Martin & Rose misconstrue the realisation relation between two levels of symbolic abstraction, context and discourse semantics, as one of composition at the same level.

[2] To be clear, the claims here are that:

• granting includes applying for
• granting includes giving
• granting includes refusing.

The logical inconsistencies here are therefore as follows:

• inclusion is a relation of extension: composition (meronymy), but
• the relation of granting to applying for is enhancement: time or cause-condition (collocation),
• the relation of granting to giving is elaboration (synonymy),
• the relation of granting to refusing is elaboration (antonymy).

Mistaking Range For Medium And Medium For Agent

Martin & Rose (2007: 92):

Amnesty is construed here as a commodity that is given or refused to various recipients, by an implicit giver (the Commission), and is also demanded by potential recipients (police officers). The central elements in this construal are the processes of exchanging (given, not given, refused, applied for), the nuclear element is the commodity exchanged (amnesty), and the marginal elements are its givers and recipients. We can represent these nuclear relations in Figure 3.10.

 

Blogger Comments:

[1] Here Martin & Rose use the terms of speech function (enacting interpersonal meaning) — exchanging (giving or demanding) commodities (goods–&–services or information) — as the basis for determining the relative involvement of participants in a process, in their rebranding of Halliday's ergative model of transitivity (construing experiential meaning) as a discourse semantic system. The confusion is thus one of metafunction.

[2] To be clear, Martin & Rose provide no argument as to why processes are central, why the commodity (Range misinterpreted as Medium) is nuclear, or why givers (Medium misinterpreted as Agent) and recipients (Beneficiaries) are marginal.  This confirms that Martin & Rose do not understand the ergative principle on which Halliday's model of clause nuclearity is based.

[3] To be clear, Martin & Rose misinterpret the ergative functions of these clauses, mistaking Range for Medium and Medium for Agent:

the Commission	may grant	amnesty	to those who plead guilty
Medium	Process	Range	Beneficiary

the Commission	does not give	amnesty	to innocent people
Medium	Process	Range	Beneficiary

the Commission	does not give	amnesty	to those who claim to be innocent
Medium	Process	Range	Beneficiary

the Commission	refused	amnesty	to the police officers
Medium	Process	Range	Beneficiary

the police officers	applied for	amnesty
Medium	Process	Range

That is, it is through the Commission and the police officers that the processes of 'granting' etc. are actualised, with amnesty as the domain of such processes.

These errors invalidate the model in Figure 3.10, even in its own terms.  In Halliday's original model, the Commission and the police officers are nuclear, with amnesty outside the nucleus:




[4] To be clear, Martin & Rose construe nuclear relations

• as lexical (and discourse semantic), even though the relations obtain between elements of grammatical structure, and
• as nuclear, even though, in their understanding, it is not the nucleus to which other elements are related.

∞

The bigger picture here is that Martin & Rose are taking Halliday's ergative model of clause grammar and rebranding it as part of Martin's model of discourse semantics.  Yet, in doing so, they demonstrate that they do not understand the principle on which Halliday's model is based, and that they cannot apply it accurately to data.