Taxonomic Relations: Theorised On A Misunderstanding Of 'Lexical Item'

Martin & Rose (2007: 81-2):

Repetition and synonymy are particularly useful resources where the field of a text is very complex. They enable us to keep one or more lexical strings relatively simple, while complex lexical relations are constructed around them. For this reason, technical texts in many fields are common contexts to find repetition and synonymy. The Reconciliation Act is one such text. Its ‘purposes’ phase is presented below with some key lexical items highlighted.

To provide for the investigation and the establishment of as complete a picture as possible of the nature, causes and extent of gross violations of human rights ... ;
the granting of amnesty to persons who make full disclosure of all the relevant facts … ;
affording victims an opportunity to relate the violations they suffered;
the taking of measures aimed at the granting of reparation ... ;
reporting to the Nation about such violations and victims;
the making of recommendations aimed at the prevention of the commission of gross violations of human rights;
and for the said purposes to provide for the establishment of a Truth and Reconciliation Commission, a Committee on Human Rights Violations, a Committee on Amnesty and a Committee on Reparation and Rehabilitation;
and to confer certain powers on, assign certain functions to and impose certain duties upon that Commission and those Committees;
and to provide for matters connected therewith.

Blogger Comments:

Leaving aside the fact that lexical items are lexicogrammatical, not discourse semantic, this demonstrates that Martin and Rose's entire model of taxonomic relations is built on a fundamental misunderstanding of what constitutes a lexical item.  In SFL theory, a lexical item is the synthetic realisation of the most delicate lexicogrammatical features.  Of the 17 highlighted elements, only 5 are genuine lexical items:

1. victims
2. violations
3. victims
4. Commission
5. Committees

The 12 that are falsely claimed to be lexical items are:

1. complete a picture
2. gross violations of human rights
3. granting of amnesty
4. full disclosure
5. relevant facts
6. violations they suffered
7. granting of reparation
8. gross violations of human rights
9. Truth and Reconciliation Commission
10. Committee on Human Rights Violations
11. Committee on Amnesty
12. Committee on Reparation and Rehabilitation

As can been seen, these are multi-word portions of (grammatical) nominal groups, all 12 of which feature multiple lexical items, and 10 of which also feature grammatical items (a, of, of, they, of, of, and, on, on, on, and).

The Notion Of Taxonomic Relations (Lexical Cohesion) Construing "Field" (Ideational Semantics)

Martin & Rose (2007: 81):

Taxonomic relations between lexical items are interpreted in terms of the field, as the reader or listener understands it. For example, a reader who is familiar with South African history would recognise the co-class relation between an Englishman and the ‘Boer’ Afrikaners, and interpret it in terms of the historical conflict between these ethnic groups. It is with this expectation of ethnic conflict that the reader interprets as remarkable the popularity of Helena’s English lover even with the ‘Boer’ Afrikaners. So taxonomic relations help construe a field of experience as a text unfolds, by building on the expectancy opened up by each lexical item, or by countering such expectancy.

Blogger Comments:

[1] This claim is falsified by the examples of taxonomic relations provided by Martin & Rose on the same page:

• repetition: marry – married – marriage
• synonyms: marriage – wedding
• antonyms: marriage – divorce
• converses: wife – husband
• scales: hot – warm – tepid – cold
• cycles: Sunday – Monday – Tuesday …
• class-member: relationship – marriage
• co-class: marriage – friendship
• whole-part: body – arms – hands
• co-part: face – hands – eyes – throat – head – brain.

Clearly, these relations are not "interpreted in terms of the field" of the text under discussion; they are, as Halliday & Matthiessen (2014: 643) put it, 'inherent in the organisation of lexis'.

[2] Here again the term 'field' is mistakenly used for 'domain', the semantic counterpart of a contextual field.

[3] To be clear, the "co-class relation" between the lexical items Englishman and Boer or Afrikaner is given in the words themselves, and so, is recognisable to English speakers in general, rather than just to the subset who are familiar with South African history.

The counter-expectancy, on the other hand, is available to the reader in the actual text (misquoted by Martin & Rose):

Even if he was an Englishman, he was popular with all the ‘Boer’ Afrikaners.

The logico-semantic relation between the two attributive clauses is concessive condition, whose meaning is if P then contrary to expectation Q:

If: he was an Englishman
then contrary to expectation: he was popular with all the ‘Boer’ Afrikaners

[4] As demonstrated above, this causal-conditional conjunctive relation (so) is unwarranted.  The conclusion does not follow from the preceding argument.

[5] To be clear, in SFL theory, what "taxonomic relations" actually do — since they are a rebranding of the textual system of lexical cohesion — is contribute to the texture of a text.

[6] In terms of SFL theory, the expression "field of experience" confuses the ideational dimension of a culture (field) with the non-semiotic realm (experience) which language transforms into meaning.  

This is compounded by the fact that Martin & Rose misconstrue field as a dimension of register, itself misconstrued as context, and compounded further still, by the fact what they refer to as field is actually, in SFL terms, the ideational domain (semantics) that realises a field (context).

[7] As demonstrated above (in [3]), the counter-expectancy that Martin & Rose attribute to taxonomic relations was actually realised in the logico-semantic relation between two clauses.

Misconstruing Ideational Semantics As Contextual Field

Martin & Rose (2007: 81):

Each lexical item in a text expects further lexical items to follow that are related to it in one of these five general ways. A lexical item initiates or expands on the field of a text, and this field expects a predictable range of related lexical items to follow.

Blogger Comments:

[1] To be clear, lexical items, as the name suggests, are located on the stratum of lexicogrammar.  Here, as throughout this chapter, lexical items are misunderstood as being located on Martin's stratum of discourse semantics, a higher level of symbolic abstraction.

[2] To be clear, in SFL theory, 'field' is the ideational dimension of context, the culture modelled as a semiotic system.  Here, and throughout, Martin & Rose unwittingly use 'field' to refer to the semantic correlate of a field — what Halliday & Matthiessen (1999: 323) term a 'domain'.  The misunderstanding is further complicated by the fact that Martin & Rose mistake 'field' for a dimension of register, misunderstood as context.

[3] In terms of SFL theory, this confuses the tendency of lexical items to co-occur — in lexical cohesion: the syntagmatic relation of collocation — with paradigmatic relations in lexical sets (repetition, synonymy, antonymy, hyponymy and meronymy); see Halliday & Matthiessen (2014: 644).  To be clear, collocation is the one type of lexical cohesion that has not been rebranded by Martin & Rose as feature of their discourse system of taxonomic relations.

[4] This misunderstands the relation between lexical item and field.  A Lexical item does not "initiate" or "expand" the "field of a text".  In SFL theory, a lexical item is the synthetic realisation of a bundle of the most delicate lexicogrammatical features, which realise semantic features, including those of ideational semantics, which realise contextual features, including those of field.  But see [2] above for what Martin & Rose misunderstand as field.

[5] To be clear, in SFL theory, the relative probabilities of lexical item instantiation in a text are a property of the lexicogrammatical systems of a register that realises a specific field.  Those probabilities rise and fall on the basis of ongoing instantiations during logogenesis.

The Taxonomic Relations System

Martin & Rose (2007: 80-1):

This range of taxonomic relations is set out in Figure 3.6.

Blogger Comments:

To be clear, the unacknowledged source of these misunderstood ideas is the lexical cohesion of Halliday & Hasan (1976), later elaborated in Halliday ± Matthiessen (1985, 1994, 2004, 2014), which is lexicogrammatical, not semantic, in terms of stratification, and textual, not experiential, in terms of metafunction.

Cf.  Halliday & Matthiessen (2014: 643-4), where the types of relation are theoretically interpreted in terms of axis (paradigmatic v syntagmatic), logical relations (elaborating v extending v enhancing) and experiential relations (identity v attribution):

The primary types of lexical relations are listed in Table 9-17. They derive from either the paradigmatic or the syntagmatic organisation of lexis. (i) The paradigmatic relations are inherent in the organisation of lexis as a resource, as represented in Roget’s Thesaurus. They can be interpreted in terms of elaboration and extension, two of the subtypes of expansion that are already familiar from the logico-semantic relations used in forming clause complexes and the corresponding conjunctive relations presented earlier in this chapter. (ii) The syntagmatic relations hold between lexical items in a syntagm that tend to occur together, or collocate with one another. Collocates of a lexical item can be found in the entries of certain modern dictionaries based on corpus investigations. Since syntagmatic organisation and paradigmatic organisation represent two different dimensions of patterning, any pair of lexical items can involve both.