David Rose On Stratification [17/8/12]

On 17 Aug  2012, David Rose wrote on the Sys-Func and Sysfling lists:

For similar reasons, perhaps we could give up the triplet 'meaning, wording, sounding' as misleading, since as Halliday has said meaning can "refer to patterns at all strata"

Then we could sensibly distinguish discourse semantics, clause semantics, and tone group semantics, recognise the semantic contributions of each stratum, and celebrate the descriptive contributions of each group of researchers.

Blogger Comments:

[1] The use of the terms 'meaning, wording, sounding' as descriptions of linguistic strata is not misleading.  They are terms that Halliday himself uses to clarify what he means by the strata that are more formally labelled as semantics, lexicogrammar and phonology.  What is misleading is the claim that they are misleading.

The distinction between meaning and wording is, of course, also made within the lexicogrammar with respect to projection.  Mental processes project meanings (ideas); verbal processes project wordings (locutions).  These processes project the semiotic order of experience: the content plane of language.  [See for example Halliday & Matthiessen (2004: 443).]

[2] Halliday's stratification model uses the more specific sense of 'meaning', as his use of the terms 'meaning, wording, sounding' suggests, and as he stated explicitly in a quote previously cited by Rose himself:

Note that “meaning” is here being used in its narrower, more specific sense, to refer just to patterns in semantics.

[3] To speak of 'discourse semantics, clause semantics, and tone group semantics' not only makes the term 'semantics' redundant, but ignores what strata represent: different levels of symbolic abstraction.  Note also that the stratum of phonology has been reduced here to one of its rank units, the tone group, to (knowingly) conceal the more obvious absurdities of, say, syllable or phoneme semantics.

[4] The 'descriptive contributions of each group of researchers' can be "celebrated" by anyone who wishes to "celebrate descriptive contributions".  Understanding the theory accurately in its own terms is the issue, not ± celebration.

Clearly, the mention of this as an issue is an example of the logical fallacy known as the argument from adverse consequences, which rejects an argument because its consequences are undesirable, or because accepting it could mean accepting something we would prefer not to acknowledge.

Jim Martin On Context, Instantiation & Stratification

On Sysfling and Sys-Func, on 17 August 2012, Jim Martin wrote:

One possibility would be to give up the terms 'context of culture' and 'context of situation' as confusing.

Genre and register (field, tenor and mode) can be used as names of strata in models using a stratified model of context.

The term context alone can be used in models with just a single stratum of context (field, tenor and mode).

My impression is that both the stratified and unstratified models of context privilege context as a stratum of meaning, moving upwards in abstraction from (discourse) semantics – that is their modelling ideal. 

In a stratified model, context as a stratum of meaning is formalised in genre networks realised through register (field, tenor and mode) networks. In an unstratified model context as a stratum of meaning is formalised as field, tenor and mode networks.

All strata instantiate, so in either model system [formalised in networks as phonology/graphology, lexicogrammar, (discourse) semantics, plus context as field, tenor and mode or as register (field, tenor and mode) plus genre] is instantiated in text. We don't give a separate name to the instantiation of phonology, or the instantiation of lexicogrammar, or the instantiation of (discourse) semantics, so why give a special name to the instantiation of the stratum of context (i.e. context of culture instantiated as context of situation) and distinguished from the collective instantiation of phonology/graphology, lexicogrammar and semantics as text? We can simply have a stratified system, with context as a stratum (or two), realised through (discourse) semantics realised through lexicogrammar realised through phonology/graphology - all instantiated as text (or readings of text if we want to push instantiation a rung further and treat text as still a meaning potential to some degree). 

We of course have to bring multimodality into the picture; but that is the same problematic issue whether we have an unstratified model of context or not. Our notion of system on the instantiation hierarchy has ultimately to be broadened to allow for coupling across modalities, each with their own realisation hierarchy, so we end up with instances of multimodal text.

So I guess I am suggesting that the notion of context as a higher level stratum of meaning doesn't seem to be properly reconciled with instantiation in SFL models that distinguish the instantiation of context of culture in context of situation from the instantiation of language systems in text. If context is a higher stratum of meaning, then there is one instantiation process, not two.

Blogger Comments:

[1] The term 'context of culture' simply means context as system, and the term 'context of situation' simply means context as instance, where 'context' means a higher (more abstract) semiotic that is realised in language [ie Halliday's stratified model].  Confusion only arises when context is construed instead as language (register and genre) [ie Martin's stratified model].  Construing context as language suggests that Martin understood context as co-text when he devised his model.

[2] If genre and register are construed as contextual strata, then the meaning (valeur) of both 'context' and 'register' changes.  'Context' is construed as language instead of context, and 'register' is construed as more abstract than semantics instead of a more specific type of language.  This means that 'register' no longer means 'a functional variety of language', since higher strata are not functional varieties of lower strata — eg lexicogrammar is not a functional variety of phonology.  Using the term 'register' for context is thus likely to create confusion.

[3] The impression that "both the stratified and unstratified models of context privilege context as a stratum of meaning" in the narrow sense of linguistic meaning is a false impression.  In Martin's stratified model, contextual strata are construed as meaning, and accordingly, to a limited extent, they can be seen as construing semantics.  However, in Halliday's unstratified model, on the other hand, context is realised in meaning (semantics).  This demonstrates the confusion that can arise from construing "all strata make meaning" (semogenesis) as "all strata have meaning".

[4] The clause "all strata instantiate" is as big a source of theoretical confusion as "all strata make meaning".  The underlying reason for this is that the relation between instance and system is attributive.  That is, in a congruent representation of the theory, the instance pole of the cline is Carrier and the system pole of the cline is Attribute.  However, the verb 'instantiate' does not function as an Attributive process, as shown by the fact that, unlike Attributive processes, it can be used in receptive clauses.  This means that the theoretical relation cannot be expressed congruently using the verb 'instantiate'.  So the clause 'all strata instantiate' is, at best, an incongruent expression of the theory, which needs to be unpacked in a way that is true to the theoretical meaning.  And this is complicated further by the fact that the theoretical notion of instantiation also refers to a process: the selection of features in system networks and the activation of their realisation statements.

[5] It is simply not true that, in both models, context "is instantiated in text".  In Halliday's model, text is an instance of the system of language, not an instance of context.  It is the context of situation — not the text — that is an  instance of the context of culture.  The text, as an instance of language, and the situation, as an instance of context, are related stratally by realisation (symbolic abstraction).

[6] One reason for using the terms 'context of culture' and 'context of situation' when talking about the cline of instantiation at the level of context is that it clarifies which pole of the cline — system or instance — we are referring to.  We do the same in the case of the linguistic strata (semantics, lexicogrammar, phonology) when we use the term 'text' to refer to the instance pole of the cline.

[7] Again, context is not 'instantiated as text' (see [5]).  Text realises context (of situation).

[8] "Readings of text" does not "push instantiation a rung further".  The relation between 'readings of a text' and a text is not the same as the (instantial) relation between a text and a linguistic system.  Locating 'readings of a text' as a point on the cline of instantiation creates a theoretical inconsistency.

[9] The text is an actualised instance of meaning potential: an ongoing instantial system (with or without 'readings of text' on the cline of instantiation — or anywhere else.)

[10] The notion of 'system' on the cline of instantiation does not need "ultimately to be broadened to allow for coupling across modalities".  The cline of instantiation models the relation between a system and instances of a system.  To the extent that "coupling" just means the co-selection of features, this is already built into the theoretical model as probabilities of co-selection at the system pole, as differences in probabilities across registers, in the middle of the cline, and as differences in actual co-selection frequencies at the instance pole.  And also to the extent that "coupling" just means the co-selection of features, it misses the point of co-selection, since what is significant about co-selected features is not that they are "coupled", but the relations between them, as defined by the architecture of the theory.

[11] It is certainly true that "the notion of context as a higher level stratum of meaning doesn't seem to be properly reconciled with instantiation in SFL models that distinguish the instantiation of context of culture in context of situation from the instantiation of language systems in text".  But the only stratification model that treats context as "a higher level stratum of meaning" is Martin's model.  In Halliday's model, meaning is located stratally in semantics.  Again, this demonstrates the confusion that can arise from construing "all strata make meaning" (semogenesis) as "all strata have meaning".

[12] Context is not a higher stratum of linguistic meaning on Halliday's model, and the process of instantiation can be viewed at any of the levels of symbolic abstraction: context, semantics, lexicogrammar, phonology.  The question of there being one or two processes of instantiation does not arise in Halliday's model.

David Rose On Register

David Rose wrote on Sysfling and Sys-Func on 17 August 2012: 

I just wanted to clarify, that in both Halliday's and Martin's interpretations, register refers to the relation between field, tenor and mode and their linguistic realisation. The difference is that Halliday defines the relation 'from below' as "the linguistic features which are typically associated with...particular values of the field, mode and tenor" (Halliday and Hasan 1976), while Martin defines it 'from above' as variations in "field, tenor and mode realised through language".

The difference seems to me not merely in terminology but in appliability, depending on one's goals. Likewise whether genre is construed as a more abstract level of context, or as a sub-class of tenor (functional tenor) or mode (rhetorical mode).

Blogger Comments:

[1] In Halliday's model, the term 'register' refers to a point of variation on the cline of instantiation.  Looked at from the system pole, each register is a subpotential of the linguistic system that realises a subpotential of the system of context; looked at from the instance pole, each register is a type of instance of the linguistic system, a text type, that realises a type of instance of the context system, a situation type.

In Martin's model, on the other hand, the term 'register' refers to a stratum.  This means it is modelled as a system — not a subpotential of the system — that is more abstract than the content plane of language. However, 'context' does not have the same meaning in Martin's model as Halliday's model, since its contextual strata, register and genre, are both levels of language rather than a semiotic that is more abstract than language.  As a result, the terms 'field' 'tenor' and 'mode' — at least to the extent that they are used self-consistently — also mean differently in Martin's model.

[2] As such, the difference between the models is not that Halliday "defines the relation 'from below'" and Martin defines it 'from above'".

[3] As such, the difference between the models is not a difference in terminology — "merely" or otherwise.  The models do not use different terms for the same theoretical meaning (valeur); they (confusingly) use the same term for different meanings (valeurs).

[4] As such, not "likewise" for genre — not that an argument for genre as a stratum of context was actually made.

David Rose On Stratification [15/8/12]

At 10:39am on 15/8/12, David Rose sent the following email to Sysfling and Sys-func entitled All strata make patterns of meaning:

For those concerned with our ongoing discussion about the nature of stratification, this recent clarification from MAKH could form a point of articulation/agreement, around his (and Firth's) 'broader sense' of meaning...
'Realization is the relationship among strata… wordings realize patterns of meaning, which we refer to as the stratum of semantics. (Note that “meaning” is here being used in its narrower, more specific sense, to refer just to patterns in semantics. The same term “meaning” is also used in a broader, more general sense to refer to patterns made at both the semantic and the lexicogrammatical stratum – the “content plane”, in Hjelmslev’s terminology. Firth used to use the term to refer to patterns at all strata, those of expression as well as those of content.)'  from Complementarities in Language (2008:14)

Blogger Comment:

Note that in modelling strata and the realisational relation between them, Halliday uses "meaning" (as he says) 'in its narrower, more specific sense, to refer just to patterns in semantics'.  He then clarifies that this usage contrasts with the other 'broader, more general sense' of the term in his linguistic  tradition.

Once again, Rose has confused the process of semogenesis, in which all strata do indeed make meaning (abstract creative material process), with the hierarchy of stratification, in which meaning is realised in wording (symbolic identity).

Robin Fawcett On The Nominal Group

On 3 Aug 2012, on sysfling John Polias wrote:

Any thoughts on whether 'kind of' in the following nominal group is a measure numerative (eg p.334 IFG3) or not. It seems to be less about delimiting than about likening (I suppose likening is also a delimiting strategy).

a kind of giant mutated axolotl

Robin Fawcett replied on the same day:

Dear John (and any others interested in this fascinating area of the lexicoframmar),

I do indeed have 'thoughts' on this matter! Two papers, in fact - but I hasten to add that it's not your fault that you haven't heard of them. 

The two papers are:

(1a) Fawcett, Robin P., 2006. ‘Establishing the grammar of “typicity” in English: an exercise in scientific inquiry.’ In Huang, Guowen., Chang, Chenguang & Dai, Fan, (eds.) Functional Linguistics as Appliable Linguistics, Guangzhou: Sun Yat-sen University Press, 159-262. Reprinted from Educational Research on Foreign Languages and Art. No. 2, 3-34 and No. 3, 71-91, Guangzhou: Guangdong Teachers College of Foreign Language and Arts, 2006. 921-52. (Also available as Fawcett 2007c.

(1b) Fawcett, Robin P., 2007a. Establishing the Grammar of ‘Typicity’ in English: an Exercise in Scientific Inquiry.COMMUNAL Working Papers No. 21. Cardiff: Computational Linguistics Unit, Cardiff University (ISSN No. 0967-0254). Reprinted from Fawcett, Robin P., 2006. Available for fawcett@cardiff.ac.uk.

(2) Fawcett, Robin P., 2007b. ‘Modelling “selection” between referents in the English nominal group: an essay in scientific inquiry in linguistics’. In Butler, Christopher S., Hidalgo Downing, R., & Lavid, J., Functional Perspectives on Grammar and Discourse: In Honour of Angela Downing, Amsterdam: John Benjamins, 165-204.

As you can see from the titles, both papers are about (i) an aspect of lexicogrammar (the phenomenon that I have termed 'typicity' in (1a and 1b) and about the broader concept of 'selection between referents' in (2) and (ii) the methodologies that I used in working on identifying (a) the full range of data that are necessarily involved in considering this topic and (b) the resulting system network, realization rules and summary of the structures generated by the outputs from those realization rules.* Your example is just one of a wide set of data that need to be considered, some with overt realizations, as in your example, and some with covert realizations, as in the ambiguous example The army have recently acquired a new tank [Is it just a single new tank, or a new type of tank?] and BP have developed a new engine oil [which is unambiguous, because the head is a 'mass' noun]. Typicity, overt or covert, occurs quite frequently in most types of text, and I have suggested the concept of a typic determiner to handle its overt realizations. This new element in the nominal group is a later development in the more general concept of 'selection between referents' in the grammar of the nominal group.

Thus, if in five of them the item five is a quantifying determiner, so too are a large number in a large number of them and two kilos in two kilos of sugar - and several other types of determiner. This is on the model of other types of determiner that are filled by a nominal group, as described in Fawcett 2007b. My concept of 'selection' is taken up and used in Mathiessen's Lexical Cartography (1995: 655), where he refers back to my earlier work, which I should perhaps therefore list here too:

Fawcett, Robin P., 1974-6/81. ‘Some proposals for systemic syntax: Parts 1, 2 and 3’. In MALS Journal 1.2, 1-15, 2.1. 43-68, 2.2 36-68. Reprinted 1981 as Some Proposals for Systemic Syntax. Pontypridd: Polytechnic of Wales (now University of Glamorgan).

I would start with Fawcett 2007b, if you have it in your university library (as I expect you do) but there is far more in Fawcett 2006 and 2007a about typicity, and about the use of corpora in resolving the question of what structure to use to model its realizations.

If you, John, or any other interested person, would like an electronic version of either Fawcett 2007a or 2007b (or both), please ask.

Robin

Robin Fawcett

Emeritus Professor of Linguistics

Cardiff University

* It is, of course, the structures that are the outputs from the use of the grammar that are what most SFL grammars in fact describe - rathe than the grammar itself. Grammars such as Halliday's Introduction to Functional Grammar and my Invitation to Systemic Functional Linguistics (Fawcett 2008) are written to provide readers with a descriptive framework for analyzing texts, which is not the same thing as a grammar (which, in SFL, consists of the system network and its associated realization rules/statements).

Blogger Comment:

An example of this type of structure is given on p333 of IFG3: a kind of owl, where a kind of is analysed as an extended Numerative, but of the category type (not measure), subcategory variety.  These are discussed on the following page.

Mick O'Donnell Mistaking A Dependent Clause For A Circumstance

Michael O'Donnell wrote on sysfling on 3/7/12:

"we have been raised to love and care" 

Actor: -

Process: material (have been raised)

Goal: We

Circumstance: "to love and care" (which is maybe Result, as Reason/Purpose don't fit).

Blogger Comments:

To be clear, O'Donnell's analysis here is:

However, systemically, the instance is agnate to 

The systemic difference here is just the finiteness of the dependent clause.

On this basis, the instance can be analysed as

David Rose On Form And Function

David Rose wrote on sys-func and sysfling on 4 June 2012:

Im not sure where a form/function duality would fit in SFL theory.

Blogger Comment:

It fits, for example, in the rank scale.  The rank scale is a compositional scale of forms.  The unit of each rank is a configuration of functional elements, each of which is realised by a unit of the rank below (except in the case of rank-shift).  For example, elements of the function structure of a clause are realised in the form of groups and phrases.

David Rose On Co-text And Semantics

David Rose wrote on Sysfling at 13:53 on 30 May 2012 about the clause complex Destroy it and man is destroyed:

There are two clauses in this sentence …

destroy it, will you?

man is destroyed, isn't he?

They are related grammatically by paratactic addition 'and'.

Any conditional relation is a discourse semantic inference, recoverable from the co-text, it is not there in the grammar of the clause complex. The grammar is not sufficient to interpret this (or any other).

Blogger Comments:

This confuses co-text with levels of symbolic abstraction (stratification).

Discourse semantics, if the stratum above lexicogrammar (wording), is a higher level of symbolic abstraction (meaning).

The co-text, on the other hand, is the text that accompanies this excerpt, and its content can be analysed at each stratum of symbolic abstraction: wording (lexicogrammar) and meaning (semantics).

It is simply not true that the meaning realised in the logical relation of the clause complex can be inferred — or is only recoverable — from the co-text: there is nothing at all in the co-text that suggests the logical meaning being incongruently worded as [extension: addition] is [enhancement: condition].

It is not that 'the grammar is not sufficient to interpret this (or any other)' but that, in SFL, all grammatical analysis involves taking a trinocular perspective, and this includes determining the meaning being realised in the wording.

See here for an analysis of the clause complex in question.

Robin Fawcett On 'Would'

On 9 May 2012, on the Sysfling list, S.S. Deol wrote:

Dear All

You may comment on the use of would in the following sentence taken from an interview of a film actor:

I was doing what the script demanded and my director would scream at me saying......

To which Robin Fawcett  replied on 9 May:

A good question!  It has a simple answer, but it is one that upsets the neat patterns of system networks that we systemicists like to draw (e.g as discussed - and exemplified - in Bache, Carl, 2008.  English Tense and Aspect in Halliday’s Systemic Functional Grammar: A Critical Appraisal and an Alternative.  London: Equinox).  …

In a systemic functional grammar, the meaning of an item such as would  is captured by the selection expression of features that have been 'chosen' in the traversal of the system network.   In the Cardiff version of SFG the meaning of would in the key clause of your text-sentence (my director would scream at me) is expressed by the following selection expression (where 'trp' stands for 'time reference point'): 

[past trp, real, long period in past, repeated].

(Here 'repeated' stands for 'repeated regularly'.)  It has no connotations of 'will', 'willingness', 'intention' etc, as other uses of would have.  Its nearest systemic relationship is with another meaning that lies outside the canonical model of 'time', i.e. used to, in which the meaning is, in terms of its semantic features:

[past trp, real, long period in past, simple].

So we can refer to the multiply repeated event in your example by saying either 

My director would scream at me. 

or, using the closely related meaning that does not restrict the meaning to repeated events, represented here by [simple]: 

My director used to scream at me.

But the first is the more precise meaning, because it expresses both 'long period' and 'repetition'.  In contrast, with a 'stative' process such as 'living' we can only say I used to live in London and not *I would live in London,  because part of the meaning of would is that the event was repeated over a relatively long period of time.  (In contrast, we CAN say I would stay with my aunt (whenever I visited London, because 'staying' lasts for a relatively short period of time.)  So the 'aspectual type' of the process of the event interacts in complex ways with 'time meanings'.

Blogger Comment:

The simple answer is that, in the SFL framework, this would functions interpersonally as the Modal operator of median usuality that realises the Finite element in the Mood of the clause as exchange.  The relevant system networks are on pages 135 (clause) and 349 (verbal group) of Halliday & Matthiessen (2004).

Mick O'Donnell On Process Types [1]

In a discussion on the email list sfl_education on 23/2/12 concerning process types, Mick O'Donnell writes:

Cases where we don't really have a category: For me, we really need a 7th category for modal lexical verbs, e.g., I was required to leave (also allowed, permitted, obliged, etc.) I find it hard to say these are auxiliary verbs.

Blogger Comments:

[1] On the SFL model, examples such as  I was required/allowed/permitted/obliged to leave do not involve an auxiliary in a verbal group.  Instead, these involve a hypotactic verbal group complex.  Halliday & Matthiessen (2004: 513):

… there is a special set that only exist as causatives, where the meaning is simply that of agency: make, cause, force require, let, allow, permit, etc.  These admit of three degrees of modulation …

[2] O'Donnell's "I find it hard to say" — like "I cannot believe" — is an example of the logical fallacy known as the argument from personal incredulity.

Mick O'Donnell On Process Types [2]

Mick O'Donnell replied on 24 February 2012 08:27:

This doesn't change my point. Halliday and Matthiessen's analysis is one analysis of these phenomena. But if we can one minute assume that it is not the case that because they say something in IFG 3, then it must be true, then we can start getting back to doing real linguistics.

As I said in my email, I believe an alternative analysis of this phenomena, where require/allow/permit, etc. are treated as full verbs, and since they have grammatical patternings very distinct to material verbs, to create a new process type for them.

Blogger Comments:

[1] O'Donnell's point has no validity because it proceeds from an inaccurate rendering of the theory in question.

[2] For example, the verbs in these verbal groups are neither 'auxiliary' nor less than "full". In clauses such as 'I was required to leave', 'was required' is the alpha verbal group and 'to leave' is the beta verbal group in a verbal group complex.

[3] The distinct grammatical patterns reflect the fact that these verbs of modulation introduce causation into the clause by means of a verbal group complex, adding the feature of agency to the clause as a whole.

[4] Theory improvement proceeds from an accurate rendering of the theory in question.  The issue is accurate interpretation, not dogmatic belief.

Register Vs Text Type

'Register' and 'text type' occupy the same point on the cline of instantiation.
Looking from the system pole, it is 'register'.
Looking from the instance pole, it is 'text type'.

Why an instance is 'a token of a type'

The organising principle of the cline of instantiation is attribution (Halliday & Matthiessen 1999: 14-5, 145).

Attribution is concerned with class membership.
Attribution + elaboration includes type–subtype relations.

But, the most delicate subtype has just a single member.
The relation between a category and its single member is identification.
The category and the member uniquely identify each other.
The member is the Token that realises the category Value.
An instance is the Token that realises the most delicate category Value.

This is the distinction between instance and subtype.

Realisation

Choices at higher strata do not determine choices at lower strata.
Function does not determine form.
Paradigm does not determine syntagm.

Choices at higher strata do not precede choices at lower strata.

Function does not precede form.

Paradigm does not precede syntagm.

Strata are different levels of symbolic abstraction.

Function and form are different levels of symbolic abstraction.

Paradigm and syntagm are different levels of symbolic abstraction.

The relation between these different levels of symbolic abstraction is 'intensive identifying', not 'circumstantial identifying' (ie neither 'causal' nor 'temporal').

Types Of Theoretical Misunderstanding

Misunderstandings of SFL theory include the following two types:

(1) Misunderstanding the terminology of the theory

(2) Misunderstanding the architecture of language proposed by the theory

A complex variant of the second type is the practice of blending two distinct dimensions while purporting to be concerned with only one of them.

David Rose On Instantiation

On 15 January 2012 at 08:31, David Rose wrote on Sysfling:

It seems to me that when reasoning about grammar, what's implicitly left to intuition is the lexical relations of the words that instantiate the grammatical categories we are interested in.

Blogger Comments:

Words do not instantiate grammatical categories.

If words — in this sense: lexical items — are the synthetic realisations of particular lexicogrammatical features (Halliday & Matthiessen 1999: 198-9), then they extend right along the cline of instantiation, from the general system, to the different probabilities of different registers, to different actualised texts (instances).

David Rose On Realisation

In a discussion on the email list sfl_education on 7/12/11, David Rose writes:

For me the issue is not meaning vs structure, since these are really two faces of the same phenomenon. 

Rather the issue is the pedagogic sequence from instance to system and vice versa. 

Since meanings and the structures that realise them only occur in texts, our approach starts with whole texts, in which learners encounter instances and gradually accumulate systems. 

Using a simple technique of guided repetition we can teach beginning learners to read and say a short text within an hour, and then use this as basis for focusing on particular features. As they acquire fragments of systems these can be used to illuminate new instances and so on.

Blogger Comments:

In SFL, meaning and structure are not "really two faces of the same phenomenon", and structures do not realise meanings.  Rather:

(1) Meaning is a stratum in the hierarchy of symbolic abstraction.  What realises meaning is wording, not structure.

(2) Structure, on the other hand, contrasts with system.  The distinction here is axial: between the syntagmatic and paradigmatic axes, respectively.  What structure realises is system, not meaning.

That is: 

meaning is realised in wording (stratification)

system is realised in structure (axis)

so:

stratification:

system on the stratum of meaning is realised in system on the stratum of wording

structure on the stratum of meaning is realised in structure on the stratum of wording

axis:

system on the stratum of meaning is realised in structure on the stratum of meaning

system on the stratum of wording is realised in structure on the stratum of wording

Sys-Func Discussion Of Theme Analysis [2]

Lise Fontaine replied on sys-func at 18:12 on 12 October 2011:

I agree with Tom's view.  There are many examples related to this:

after they got married

right after they got married

one year after they got married

10 years after the got married

and there are other prepositions which follow similar patterns.  In IFG terms this ('one year after') constitutes a prepositional group within a prepositional phrase.

Blogger Comments:

(1) Lise says she agrees with Tom's view, even though Tom mistakenly analyses one year after as a conjunction group, while she mistakenly analyses it as a preposition group.

(2) In IFG terms:

after they got married is a clause initiated by the (unmodified) conjunction group after

right after they got married is a clause initiated by the (premodified) conjunction group right after

one year [[after they got married]] is a nominal group with an embedded clause as Qualifier

10 years [[after they got married]] is a nominal group with an embedded clause as Qualifier
and
those twenty unforgettable though nightmarish years [[after they got married]] is a nominal group with an embedded clause as Qualifier

Sys-Func Discussion Of Theme Analysis [3]

Lise Fontaine continued on sys-func at 18:52 on 13 October 2011:

I'm just leaving for a conference so my reply won't be as detailed as I would like. I should say though that I wasn't suggesting I was representing the views of IFG in my reply, only that I agreed with Tom's analysis.

Blogger comments:

Lise says "I wasn't suggesting I was representing the views of IFG in my reply", but in that reply she had written:

"In IFG terms this ('one year after') constitutes a prepositional group within a prepositional phrase".

1. David Rose On Stratification & Metaredundancy

David Rose firstly writes in response to my comments on Jim Martin On Stratification:

contextual strata are metaredundant on the redundant relation between discourse semantics and lexicogrammar... hence relating clause meanings directly to context skips a layer of metaredundancy

Blogger Comments:

(1) clause meanings

The meanings of clauses are semantic not lexicogrammatical. Hence relating clause meanings directly to context does not skip a stratum.

(2) layer of metaredundancy

Strata are not layers of metaredundancy. Metaredundancy is a redundancy on a redundancy. Semantics is metaredundant on the redundant relation between lexicogrammar and phonology. Semantics, lexicogrammar and phonology are not "layers of metaredundancy".

2. David Rose On Realisation

David Rose secondly writes in response to my comments on Jim Martin on Stratification:

clause meanings are realised by system-structure relations cycling through grammatical ranks i.e. meanings are realised by relations between elements of structures in texts, instantiating relations between features in systems, at each rank within phonology, grammar and discourse... (no need to do it twice)

Blogger Comments:

(1) clause meanings are realised by system-structure relations; meanings are realised by relations between elements of structures

This confuses the distinction between stratification and axis. In terms of stratification, meanings (semantics) are realised by wordings (lexicogrammar) — whatever the axis (paradigmatic or syntagmatic). In terms of axis, (paradigmatic) systems are realised by (syntagmatic) structures — whatever the stratum.

(2) meanings are realised by relations between elements of structures in texts

'In texts' refers only to the instance pole of the cline of instantiation. The realisation of meaning (semantics) in wording (lexicogrammar) applies to the whole cline, not just the instance. Hence: "stuck in the instance".

3. David Rose On Instantiation

David Rose thirdly writes in response to my comments on Jim Martin on Stratification:

discourse semantic systems are the potential for co-textualising grammatical meanings, instantiated as discourse semantic structures in text

Blogger Comment:

Instantiation is the relation of the system to the instance (of the system). The relation between system and structure — and between semantics and lexicogrammar — is realisation.

4. David Rose On Instantiation

David Rose writes, in response to my comments on 3. David Rose On Instantiation:

All true but doesn't falsify DR's statement. How else are systems instantiated other than as structures in texts? DR's statement corrected CC's claim that "Co-textualised" meanings are only at the instance pole of the cline of instantiation — in text

My Comments:

[1] This confuses the cline of instantiation with the realisation relation between the paradigmatic and syntagmatic axes. The theoretical dimension of instantiation is the (intensive attributive) relation between the system and the instance (of the system). The relation between the paradigmatic axis (systems) and the syntagmatic axis (structures) is realisation (an intensive identifying relation).

Put simply: a structure is not an instance of a system.

[2] By definition, "co-textualised" meanings are meanings of the text. The text is at the instance pole of the cline of instantiation.

5. David Rose On Meaning

David Rose writes in response to my comments on 2. David Rose On Realisation:

Again this reduction does not falsify DR's statement. SFL is a relational theory of meaning, in which meanings are realised by the whole set of relations in language, including stratal, axial and rank, which DR's statement distilled and explicated. Meaning making is not reducible to a one-dimensional relation between strata.

Blogger Comments:

(1) Again this reduction does not falsify DR's statement.

A distinction — in this case between stratification and axis — is not a reduction.

(2) meanings are realised by the whole set of relations in language, including stratal, axial and rank

The use of 'meaning' here ignores theoretical distinctions. Meanings are realised by wordings (stratification); systems are realised by structures (axis); the functional elements of a higher rank are realised by units of the rank below.

(3) which DR's statement distilled and explicated

Ignoring theoretical distinctions is neither distillation nor explication.

(4) Meaning making is not reducible to a one-dimensional relation between strata.

The original post was about the realisation of meaning, not meaning making. The realisation of meaning is a stratal relation between semantics and lexicogrammar (the axial relation of realisation on the semantic stratum is between system and structure).

(5) Meaning making is not reducible to a one-dimensional relation between strata.

That claim was not made. This is the logical fallacy known as 'The Straw Man'.

http://www.logicalfallacies.info/ambiguity/straw-man/

6. David Rose On The Cline Of Instantiation

David Rose writes in response to my comments on 2. David Rose On Realisation: 

Yes, and 'relations between features in systems' refers to the system pole of the cline. Relations between elements of structure are dimensions of meaning making in SFL theory, whether viewed from structural potentials realising features in systems or actual structures instantiated in texts

Blogger Comments:

[1] This misunderstands the cline of instantiation. The cline of instantiation is (an intensive attributive) relation between the system and instances (of the system). There are no (syntagmatic) "relations between elements of structure" at the (paradigmatic) system pole of the cline of instantiation.

[2] This misunderstands the cline of instantiation and merges it with (axial) realisation. The cline of instantiation is (an intensive attributive) relation between the system and instances (of the system). "Actual structures" are not instances of a system. The relation between system and structure is realisation (an intensive identifying relation).

7. David Rose On Stratification

David Rose writes in response to my comments on 1. David Rose On Stratification & Metaredundancy:

The meanings of clauses are made by lexicogrammatical systems/structures. They do not include meanings made by discourse semantic systems/structures, which co-textualise (redound on) these lexicogrammatical meanings. Hence relating clause meanings directly to context skips a layer of (OK) redundancy.

Blogger Comments:

(1) The meanings of clauses are made by lexicogrammatical systems/structures.

This misunderstands stratification. The relation between meaning (semantic stratum) and wording (lexicogrammatical stratum) is realisation — an intensive identifying process, not a creative abstract material process ('made by'). Wording (eg clauses) realises meaning.

(2) They do not include meanings made by discourse semantic systems/structures, which co-textualise (redound on) these lexicogrammatical meanings. Hence relating clause meanings directly to context skips a layer of (OK) redundancy.

This misunderstands stratification. Meaning is the level of symbolic abstraction modelled as the semantic stratum. The meanings realised by wordings — including wordings such as those of clause rank and of cohesion — are at the semantic level of symbolic abstraction. No stratum is "skipped".

Robin Fawcett On Stratal Relations [1]

In a discussion on Sysfling [29/6/11], Robin Fawcett opined:

Indeed, if Michael Halliday and Christian Matthiessen had formed a clear view of the way in which the choices described in their Construing Experience through Meaning determine the choices in the major system networks of the lexicogrammar, they would surely have said so in that book. I have looked hard for a section that makes this connection, but I have yet to find it. This suggests that the model proposed there is simply one possible, half-complete hypothesis that needs to be subject to the normal process in science of development, testing, evaluation, revision (or rejection), retesting, re-evaluation, and so on.

Blogger Comments:

(1) Here are some of the quotes that Robin Fawcett was unable to find.

Halliday & Matthiessen (1999: 375):

More specifically, inter-stratal realisation is specified by means of inter-stratal preselection: contextual features are realised by preselection within the semantic system, semantic features are realised by preselection within the lexicogrammatical system, and lexicogrammatical features are realised by preselection within the phonological/graphological system.

Halliday & Matthiessen (1999: 378):

The realisational relationship between semantics and lexicogrammar is one of preselection: semantic features … are realised in lexicogrammar by means of prespecification of lexicogrammatical information, most centrally preselection of lexicogrammatical features.

(2) Regarding the notion that strata "determine" one another, note that higher stratal choices do not cause lower stratal choices.

Halliday & Matthiessen (1999: 25):

In any stratal system (i.e. any system where there are two strata such that one is the realisation of the other) there is no temporal or causal ordering between the strata. … the relationship is an intensive one, not a causal circumstantial one.

(3) On realisation as an analogue of cause:effect in classical physics:

Halliday & Matthiessen (1999: 510):

… we do not yet fully understand the nature of the relationship that is the semiotic analogue of the “cause : effect” of classical physics: this is the problem of realisation.

Having had his claim falsified, Fawcett subsequently re-interpreted his original question (extending it) in an attempt to obscure the falsification.  This is a subtype of the logical fallacy known as the 'red herring':

The red herring is as much a debate tactic as it is a logical fallacy. It is a fallacy of distraction, and is committed when a listener attempts to divert an arguer from his argument by introducing another topic.

http://www.logicalfallacies.info/relevance/red-herring/

Robin Fawcett On Stratal Relations [2]

In a discussion on Sysfling [1/7/11], Robin Fawcett reacted to the three quotes [in Part 1] from Halliday & Matthiessen (1999: 375, 378, 25) as follows:

In the first two passages we are given a picture of 'preselection' as 'predetermination', and so as something approaching 'causation'. Then in [the third] we are told that the relationship is not like that. But we are not told WHY it is not. We are offered instead a metaphor (taken from the terms of the Sydney Grammar itself), but not an explanation as to WHY the relationship is an 'intensive' one. 

The reader of these three passages will surely ask: How can it be true that choices made at one stratum are realized in choices at a lower stratum, without this being some type of 'pre-determination'? Passage [3] presents the relationship as a great mystery!

Blogger Comments:

Among other things, Fawcett does not understand that:

1. the theoretical dimension of stratification is organised on the principle of intensive identification — it is not a metaphor, it is what stratification means in this way of modelling language: higher level Value is realised by lower level Token;
2. the strata are thus different levels of symbolic abstraction of the same phenomenon (language) — there can be no chain of command across different levels of symbolic abstraction;
3. semantic features being realised by the 'preselection' of lexicogrammatical features simply means that features at the higher level of abstraction entail features at the lower level of symbolic abstraction — same phenomenon, different levels of abstraction.  For example, the logico-semantic relation of 'cause' can be realised incongruently by the preselection of lexicogrammatical features relating to participants, processes or circumstances at clause rank.

Halliday & Matthiessen (1999: 94) anticipated Fawcett's misinterpretation (in the very same text):

Such selections have been referred to as “pre-selections”, but in order to avoid any connotations of temporal sequence, we prefer the term “selection” for such relations in the ideation base.