David Rose Misunderstanding Delicacy And Instantiation

David Rose replied to Beatriz Quiroz on Sysfling on 9 Sept 2023 at 18:50:

The more delicate a system, the more instantial it is, so the more it couples features from multiple systems, eg DEIXIS and IDENTIFICATION

The IFG discussion of DEIXIS is interesting as it shifts back and forth between LG and DS

Blogger Comments:

[1] This is misleading because it is untrue. It is the whole system, along the dimension of delicacy, that is located on the cline of instantiation from potential to subpotential/instance type to instance. Clearly, the more delicate system of attributive and identifying process types, for example, is not more "instantial" than the system that includes material and mental process types. Halliday & Matthiessen (2014: 219).

Halliday & Matthiessen (1999: 327):

Instantiation is the relation between the system and the instance. When we shift attention along this scale, we are moving between the potential that is embodied in any stratum and the deployment of that potential in instances on the same stratum (between the climate and the weather, to use the analogy from our illustration below). Again, this move can be made at any degree of delicacy. 

Delicacy is the relation between the most general features and the most specific. When we shift attention from, say, 'recreation' to 'hockey' at the level (stratum) of context, or from 'syllable' to long open nasal syllable' to /pã:/ in phonology, we are moving in delicacy. Again, we can do this at any point along the instantiation scale.

[2] This is misleading because it is untrue. The number of co-selected ("coupled") features from different systems does not increase down the cline of instantiation. What is true is that the co-selection of features differs at points of variation along the cline of instantiation. That is, registers (subpotentials) differ from one another in which features are co-selected; texts (instances) differ from one another in which features are co-selected.

[3] To be clear, DEIXIS is a system of the nominal group, whereas IDENTIFICATION is Martin's misunderstanding of REFERENCE — Halliday & Hasan's system of grammatical cohesion — relocated to his stratum of discourse semantics and rebranded as his own system. Evidence here.

[4] This is misleading because it is untrue. Unsurprisingly, the discussion of DEIXIS in IFG is only concerned with Halliday's lexicogrammar, not Martin's discourse semantics. This deception is merely an attempt to legitimate Martin's model by incorporating it into Halliday's exposition of the grammar.

Beatriz Quiroz On The Nominal Group System Of Deixis

Beatriz Enriqueta Quiroz Olivares replied to David Rose on SYSFLING on 9 Sept 2023 at 11:29:

So you say in IFG4 [selective] and [non-selective] may not be LG choices? And that they are instead be comparable to ET [non-directed] and [directed] respectively? It does make sense, but I wonder... Because Table 6-3 in the same section of IFG4 shows non-specific determiners and that table does include [selective] and [non-selective] for [non-specific: partial] - something that is inconsistent with the system network shown on p. 366.

And on 10 Sept 2023 at 00:12: 

And yes, the kind of DS semantic discussion in IFG on determiners is something that most non-SFL grammarians at this point take into account when looking at determiners, at least in Spanish.

 

Blogger Comments:

[1] To be clear, this does not make sense, since features in lexicogrammatical systems, like those of the nominal group are, by definition, are lexicogrammatical, not semantic.

[2] This is misleading, because it is not true. There is no inconsistency, since Table 6-3 merely provides features that specify non-specific DEIXIS and which are not provided for that system in Figure 6-2. The feature 'selective' just means there are more delicate selections to be made, and so can appear anywhere in a network where such a distinction needs to be made.

Halliday & Matthiessen (2014: 366, 368):

[3] To be clear, there is no discourse semantic discussion in IFG on determiners or on anything else, because IFG, as the name suggests, is an introduction to Halliday's functional grammar, not an introduction to Martin's discourse semantics.

David Rose Misconstruing Discourse Semantic Systems As More Delicate Grammatical Systems

David Rose replied to Beatriz Enriqueta Quiroz Olivares on Sysfling on 9/9/23 at 9:32:

Which brings us to IFG Fig.6-2 The nominal group system network: determination

 

Deixis articulates lexicogrammar with discourse semantics, but I’m not sure this network distinguishes LG and DS criteria. The personal/demonstrative and determinative/interrogative systems are clearly LG, but more delicate systems are DS - interactant/non-interactant and selective/non-selective (ET’s directed/undirected).

That may explain its fuzziness of selective/non-selective criteria (system not named)?

How to represent more clearly??

PS In Australian lgs, personal Deictics precede Thing and demonstrative Deictics follow Thing. 

Blogger Comments:

[1] To be clear, this metaphor is theoretical nonsense. DEIXIS is a lexicogrammatical system of the nominal group; lexicogrammar and semantics are different levels of symbolic abstraction.

[2] To be clear, this is theoretical nonsense. A grammatical network does not "distinguish discourse semantic criteria" because it models grammatical potential, not semantic potential.

[3] To be clear, this is theoretical nonsense. The more delicate systems of a grammatical systems are, unsurprisingly, grammatical, not semantic, since semantic systems are on the semantic stratum, not the grammatical stratum.

[4] To be clear, the feature 'selective' merely provides the entry condition for more delicate selections, in number and proximity, whereas the feature 'non-selective' does not.

[5] To be clear, this claim is contradicted by Rose's previously posted analysis of his own data, where the demonstrative nani 'this' precedes the Thing daluk 'woman':

David Rose Mistaking Realisation Rules For Function Structures And Paradigmatic Features For Syntagmatic Units

David Rose replied to John Bateman on Sysfling on 30/8/23 at 22:41:

Just some parameters...

In a rank scale, the initial entry condition for systems is a syntagmatic unit eg clause. Their features e.g. [indicative] are realised by function structures at that rank ↘︎S^F, and each function within that structure e.g. Subject is realised by a syntagm at the next rank ↘︎nom gp.

Phonemes are syntagmatic units at the base of a rank scale. Features at syllable rank are realised by function structures eg ↘︎(Onset)^Rhyme, and each function is realised by a phoneme (complex). Phonemes are produced by tongue and lip postures with voicing, but these lie outside language in the biology of the vocal tract... Jim’s materiality (the ‘etics’ materialising ‘emics’)

So yes, it seems adult social semiosis minimally requires two syntagmatic ranks

Im interested in how that emerges from unranked protolg

Blogger Comments:

[1] This is misleading because it is untrue. The entry conditions to paradigmatic systems are paradigmatic features, not syntagmatic units. The rank scale construes formal constituents as a system of features, with the selection of a feature satisfying the entry condition to the systems of that rank.

[2] This is very misleading because it is a serious misunderstanding. Features are not realised by function structures. Importantly, in Rose's first example, ↘︎S^F is not a function structure, but a realisation statement that specifies 'Subject before Finite'. Moreover, it is activated by the selection of the feature 'declarative', not 'indicative'. The feature 'indicative' activates the insertion of a Mood element (Subject and Finite). See Halliday & Matthiessen (2014: 162).

In Rose's second example, ↘︎(Onset)^Rhyme is not a function structure, but a realisation statement of an unidentified feature that specifies 'optional Onset before Rhyme'.

[3] To be clear, the systems that specify phonemes are those that 'label tongue and lip postures with voicing' (and more), so this is not the distinction between phonology and phonetics. The difference is stratal: phonology is a higher level of symbolic abstraction than phonetics. Moreover, the materiality of phonetics is hardly Martin's discovery.

[4] Clearly, SFL Theory maintains that four ranks are required to model formal constituency in lexicogrammar and phonology. But rank only applies to the expression plane of semiotic systems without a lexicogrammar.

John Bateman On The Necessity Of Rank Across Semiotic Systems

John Bateman replied to David Rose on Sysfling on 28/8/23 at 21:03: 

picking up on [your]:

Not sure about necessity of ranks, since lowest lg ranks have only next stratum to operate in

this is an interesting modelling issue. Formally, I would tend to want to include a category even if it is not expanded if it is a logical consequence of other components of the theory. I take rank in this sense — i.e. without some unit of structure there is no place for axis to start its magic.

Also wrt Jim's :

“As systemicists interested in genesis, should we be proposing a major component of our cartography (i.e. a rank, a metafunction or a stratum) in the absence of a distinctive system of valeur?” [Martin 2010]

maybe one ranked unit is enough?

In any case, if I had to put my finger on one reoccurring (quite serious) problem with many many multimodal analyses that I see (including some waving the SFS flag), it would be lack of attention to rank — mostly signalled by system networks hanging wonderously in the air and curious 'multiple options' (e.g. 'either-and' systems) going along with that, mostly, I suspect, because the units aren't clear.

So, I'd be very hesitant to leave out rank.

Blogger Comments:

[1] See David Rose On Ranks, Axis And Strata In Protolanguage.

[2] This is misleading, because it is not true. Firstly, rank is a means of modelling formal constituency, so it only applies to strata that feature formal constituents. In language, these are lexicogrammar on the content plane and phonology and graphology on the expression plane. Since other semiotic systems do not have a lexicogrammar — they cannot be read aloud — rank can only apply to their expression plane. Rank is a local dimension, not a global dimension (Halliday & Matthiessen 2014: 32).

Secondly, the SFL model of ideational semantics demonstrates that rank is unnecessary, since it has axis, system and structure, but no rank. Instead of a rank scale of forms, there are types of phenomenon: sequence, figure and element (Halliday & Matthiessen 1999: 49), with each type being the entry condition to systems.

[3] See David Rose On Ranks, Axis And Strata In Protolanguage.

[4] To be clear, the problem here not the absence of rank, but the absence of entry conditions. These are provided by rank units in lexicogrammar and phonology/graphology, but there are no rank units on the content plane of a semiotic system other than language, because only language has a lexicogrammar.

[5] See [2] and [4].

David Rose Misunderstanding Immanence, Transcendence And 'Language Turned Back On Itself'

David Rose replied to Kieran McGillicuddy on Sysfling on 28/8/23 at 14:13:

There’s more to Firth’s apparent disavowal than meets the eye. And times have changed. Here’s a prosaic gloss...

1 Our schematic constructs must be judged with reference to their combined tool power in a way our dealings with linguistic events in the social process.

‘Our metalanguage is designed for research’

2 Such constructs have no ontological status

‘Its terms shouldn't be mistaken for reality’

3 and we do not project them as having being or existence.

‘We don’t pretend they are’

4 They are neither immanent nor transcendent,

I dunno... hence my q to Ed

5 but just language turned back on itself.

‘but now I’m just going to bend your mind’.

and added at 19:53:

Please forgive me all, my ignorance

Immanence and transcendence were of course, not currents in lx theory, but in Christian theology, of the spirit located within or above the material world. But Firth’s readers would all have known that.

 

Blogger Comments:

[1] Again, Firth's first point is that theory must be judged according to its explanatory power in modelling language in context.

[2] To be clear, Firth's second and third points are that, in his view, the schematic constructs themselves have no ontological status: no being or existence.

[3] To be clear Firth's fourth point makes reference to opposing orientations to meaning in linguistic traditions, not Christian theology. Halliday & Matthiessen (1999: 415, 416):

We can identify two main traditions in Western thinking about meaning (see Halliday, 1977):

(i) one oriented towards logic and philosophy, with language seen as a system of rules;

(ii) one oriented towards rhetoric and ethnography, with language seen as resource. …

The two orientations towards meaning thus differ externally in what disciplines they recognise as models. These external differences are associated with internal differences as well.

(i) First, the orientations differ with respect to where they locate meaning in relation to the stratal interpretation of language:

(a) intra-stratal: meaning is seen as immanent — something that is constructed in, and so is part of, language itself. The immanent interpretation of meaning is characteristic of the rhetorical-ethnographic orientation, including our own approach.

(b) extra-stratal: meaning is seen as transcendent — something that lies outside the limits of language. The transcendent interpretation of meaning is characteristic of the logico-philosophical orientation.

Many traditional notions of meaning are of the second kind — meaning as reference, meaning as idea or concept, meaning as image. These notions have in common that they are 'external' conceptions of meaning; instead of accounting for meaning in terms of a stratum within language, they interpret it in terms of some system outside of language, either the 'real world' or another semiotic system such as that of imagery.

[4] Again, Firth's fifth point is simply that linguistic theory is the use of language to model language; linguistic theory is language about language: a metalanguage.

David Rose Misunderstanding Firth

David Rose replied to Ed McDonald On Sysfling on 28/8/23 at 10:34:

Firth’s famous disavowal begs more discussion...

1 Our schematic constructs must be judged with reference to their combined tool power in a way our dealings with linguistic events in the social process.
2 Such constructs have no ontological status
3 and we do not project them as having being or existence.
4 They are neither immanent nor transcendent,
5 but just language turned back on itself. (Firth 1957 [1950]: 181)

The first clause is the argument for an appliable lx

While the 2nd clause is unarguable (they're constructs)...

3 contradicts another position on SFL theory, that its architecture is designed to mirror the phenomena it construes

4 seems unarguable but lx history q... what currents of the time is he countering?

5 foreshadows a mimetic meta language despite his mere modesty

Blogger Comments:

[1] To be clear, Rose's gloss misunderstands Firth. Firth's first point is that theory must be judged according to its explanatory power in modelling language in context.

[2] To be clear, Rose's gloss is misleading, because the claim that theoretical constructs have no ontological status is arguable, since it depends crucially on what is encompassed by the meaning of 'ontology'. 

[3] To be clear, Rose's gloss has nothing to do with Firth's point. Firth's third point follows from his second point: having no ontological status entails having no being or existence. So it doesn't contradict Rose's misunderstanding that a theory mirrors the theory it construes (for which, see the previous post).

[4] To be clear, again Rose's gloss has nothing to do with Firth's point. Firth's fourth point is that theoretical constructs are outside the distinction between meaning as immanent and meaning as transcendent. This is at odds with SFL Theory, where all meaning is taken to be immanent, including meaning used to model meaning. At this point Rose is totally lost.

[5] To be clear, Firth's fifth point is simply that linguistic theory is the use of language to model language; linguistic theory is language about language: a metalanguage.

David Rose On Theory Mirroring And Representing A Phenomenon

David Rose replied to Ed McDonald on Sysfling on 27/8/23 at 14:38:

My interest in axis doesn't in any way diminish all you say here. There’s no value judgement as to what part of the theory is more important or primal than any other. MAKH developed it initially as a research tool to map functions and structures, discovering metafunctions along the way, and an empirical model of ranks in grammar and phonology. But the theoretical formalism is not just a tool, it mirrors the organisation of the phenomenon it represents.

My interest is particularly in how it works as a configuration of abstraction and generalisation to make semiosis possible. There is actually no such thing as systems and structures. They are merely generalisations from recurrent instances of sounds and other perceptible phenomena. Yet they turn matter into meaning.

Blogger Comments:

[1] To be clear, Rose had written on 23 August 2023 at 13:55:

Can I suggest when reading the various philosophers of language cited in this thread, to remember the one missing puzzle piece that lies at the heart of systemic functional semiotics...axis.

 on 24 August 2023 at 00:01:

Axis is the systemic in systemic functional. In this theory, the coupling of generalisation with abstraction in the axial relation is the basis of semiosis.

and on 25 August 2023 at 8:40:

My original point was that axis is the basis of semiosis and that these are all dimensions in which axis proliferates in lg.

[2] To be clear, on the SFL model, a theory neither mirrors nor represents a phenomenon. If a theory mirrored its phenomenon, then there would be only one possible theory. If a theory (Token) represented a phenomenon (Value), then the phenomenon would be more abstract than the theory.

Instead, a phenomenon is a construal of first-order experience as meaning, and a theory is a metaphenomenon: a reconstrual of that first-order meaning as second-order meaning. The complication with language is that it is itself a theory of experience, a metaphenomenon, so that a theory of language is a reconstrual of that theory of experience as a theory of a theory of experience, that is: a metatheory of experience. It is because linguistic theory is 'meta' to language in this way that it is termed a 'metalanguage'.

[3] As previously explained, systems and structures are abstract things in the metalanguages that theorise languages.

David Rose On Ranks, Axis And Strata In Protolanguage

David Rose replied to John Bateman on Sysfling on 27/8/23 at 7:14:

Not sure about necessity of ranks, since lowest lg ranks have only next stratum to operate in. Morphemes and phonemes aren’t configurations, are they? Protolg is not ranked, which may relate to Jim’s qu about materiality. Which reminds me of his 2010 intro in Making Meaning Matter, which contrasted two interpretations of axis and strata in protolg...

First is Xian following MAKH... “The content-expression organisation of protolanguage is thus simultaneously stratal and axial… The dimensions of stratification and axis are not independently variable at this stage: the content stratum is organised paradigmatically, while the expression stratum is an inventory of indivisible postural or gestural syntagms.” [Matthiessen 2007: 516]

Second is JRM... “As systemicists interested in genesis, should we be proposing a major component of our cartography (i.e. a rank, a metafunction or a stratum) in the absence of a distinctive system of valeur?” [Martin 2010]

Here’s MAKH [2004/2012] on Nigel @ 15 months reinterpreted by me as axis and strata...

Blogger Comments:

[1] Clearly, if this were true, it would be an argument against the necessity of rank in language (morpheme, phoneme). To be clear, rank is one way to model formal constituency, so the lowest rank, by definition, has no constituents. But a syntagm of lowest rank units realises the structural configuration of the higher rank: a syntagm of morphemes realises word structure; a syntagm of phonemes realises syllable structure. Moreover, as a rank unit of lexicogrammar, a morpheme does not "operate in" phonology, since lexicogrammar and phonology are different levels of symbolic abstraction.

[2] To be clear, the reason why rank does not apply to protolanguage is semiotic, not material. Protolanguage has no formal constituency, and so it has nothing to model as a rank scale.

[3] To be clear, Martin's model of protolanguage posits structure where there is none. A structure is the relationships between functional elements, but there is no configuration of functions in protolanguage. Each realisation is an individual and indivisible sound or gesture.

[4] Strictly speaking, these are not syntagms, since a syntagm is a sequence of classes, and the expression stratum of protolanguage involves no sequencing.

[5] To be clear, Martin has himself violated his own principle by proposing "a major component of our cartography", structure, "in the absence of a distinctive system of valeur" that specifies structural configurations.

[6] To be clear, Rose's reinterpretation of Halliday (2004/2012) 

• proposes structure where Halliday says there is none,
• misrepresents expression (an:a, dada) as content, and 
• mistakes the resonance feature 'nasal' for a type of (oral) closure.

John Bateman On Dimensions Across Semiotic Modes

John Bateman replied to David Rose on Sysfling on 26/8/23 at 20:16:

… But the following [from Rose] is very nice to hear:

Metafunctionality as I tried to suggest, is a property of lg that its syntagmatic structuring has evolved to enable, together with ranks and strata. So it’s basic to lg but not to semiotic modes in general.

I'm wondering how well that statement goes down more generally: consequences should actually follow.

I think the question of what dimensions are common across semiotic modes is very interesting and also central for a semiotics. I'd readily take rank to have to be present, otherwise axis has no domain (s) to operate in. Structural configurations have to be configurations of something. Stratification is also a good bet. Moreover, since semiosis is inherently dynamic, we'll have instantiation and sedimentation relating to reservoires and repositories I'd expect. These would all seem to be necessarily bound up in a single 'definition' of a particular kind of signs for my money. This means there is no 'picking and choosing' to do. …

Metafunction again turns up there as a bit of a swamp... the original idea of "distinctive regions of relatively interdependent systems" gets tricky without sufficiently strongly motivated systems. And I am not yet sure to what extent the different kinds of structure (which I like a lot) actually match well with kinds of functions parallel to linguistically motivated metafunctions in nonverbal semiotic modes.

Blogger Comments:

[1] To be clear, this is not "very nice to hear" from a theoretical point of view. See David Rose Misunderstanding Metafunctionality And Axis (Inter Alia).

[2] To be clear, rank is one way of modelling formal constituency. In language, rank is a dimension of lexicogrammar and phonology/graphology. Therefore, in modelling semiotic systems without a grammar — that is, those other than language — rank can only be a dimension of the expression plane, not the content plane.

[3] This is misleading because it is untrue. To be clear, semantics has no rank scale because it has no forms to rank, and yet semantics is modelled with both axes, paradigmatic and syntagmatic; see Halliday & Matthiessen (1999).

Bateman's false claim derives from two theoretical misunderstandings. Firstly, rank is not the domain of axis. Axis and rank are distinct local dimensions. Axis has two orders: the paradigmatic and syntagmatic. Rank has four orders: clause, group/phrase, word, morpheme in lexicogrammar. Halliday & Matthiessen (2014: 32):

Secondly, structures are configurations of functions (Senser^Process) not rankscale forms (nominal group ^ verbal group). The sequencing of formal units is a syntagm.

[4] To be clear, without the stratification of content and expression, there can be no semiotic system. Semiosis requires that a Token means a Value. However, the stratification of the content plane is restricted to language, as demonstrated by the fact that only language can be read aloud, since the locutions that a Sayer projects are the wordings of the lexicogrammatical stratum.

[5] This is possibly a garbled reference to 'reservoir' and 'repertoire' in Martin's Bernstein-derived model of individuation.

[6] To be clear, this misrepresents the metafunctions. The metafunctions are "strongly motivated" as the basic types of meaning: construing experience as meaning, enacting intersubjective relations as meaning, and weaving these together as text. Each of these understandings of the metafunctions can be applied to semiotic systems. Indeed, Halliday & Matthiessen (1999: 532-3) are sceptical of there being full semiotic systems without metafunctions:

These three "metafunctions" are interdependent; no one could be developed except in the context of the other two. … There are — or could be — semiotics that are monofunctional in this way; but only very partial ones, dedicated to specific tasks. A general, all-purpose semiotic system could not evolve except in the interplay of action and reflection, a mode of understanding and a mode of doing — with itself included within its operational domain.

Jim Martin Spruiking His Group's Work On Systemic Functional Semiotics

Jim Martin wrote to Sysfling on 25/8/23 at 11:43:

Cute, but let’s not be too cavalier about dismissing work that can reasonably be counted as SFS (such as that noted below). This work is explicit enough and comprehensive enough with respect to a range of semiotic modes that probing questions can be raised (and have been), including…

• Has enough attention been paid to materiality (the ‘etics’ materialising ‘emics’)?
• Has enough attention been paid to ‘constituency’ (ranks and taxis)?
• Has enough attention been paid to the interdependency of systems and their possible correlation with types of structure (re ‘metafunctions’)?
• Has enough attention been paid to stratification (is one stratum being made to do too much work)?
• Has enough attention been paid to instantiation (especially the modelling of abductive inferencing as texts unfold)?

The answer to each of these questions is some degree of “no”; in some cases extant SFL theory makes room for improvement (cf. van Leeuwen on the ‘materiality’ of sound or colour) and in other cases it is sorely lacking (e.g. how to model coupling, commitment and the abductive inferencing highlighted in the dynamic take on discourse semantics propagated by John and his colleagues).

The paper attached from the 2010 UTS ICOM meeting touches on many of the issues raised in this thread:

Multimodal semiotics: theoretical challenges. S Dreyfus, S Hood & M Stenglin [Eds.] Semiotic Margins: meaning in multimodalities. London: Continuum. 2011. 243-270.

images

Kress & van Leeuwen Reading Images

picture books

Painter et al. Reading Visual Narratives

animations

He ‘Towards a stratified metafunctional model of animation’ Semiotica 2021; 239; 1-35

paralanguage

Ngo et al. Modelling paralanguage using SFS

emoji

Zappavigna & Logi Emoji and Social Media Paralanguage (in press CUP)

infographics

Martin & Unsworth Reading Images for Knowledge Building

space

McMurtrie The Semiotics of Movement in Space

math symbolism

Doran The Discourse of Physics

chemistry symbolism

Yu Multimodal Knowledge Building in Secondary School Chemistry Textbooks (forthcoming Bloomsbury)

Blogger Comments:

[1] To be clear, Martin's concern here is not with whether theory has been understood and consistently applied in such work, but simply with whether the work is explicit and comprehensive enough to ask questions about what else needs to be done. The academic quality of the work is immaterial.

[2] To be clear, coupling and commitment are two of Martin's misunderstandings of SFL Theory. The notion of coupling derives from Martin's misunderstanding of metafunctions and strata as interacting modules (Martin 1992: 390ff). In fact, it merely makes the banal observation that features are selected with other features in instantiation. 

The notion of commitment derives from Martin's misunderstanding of delicacy and instantiation, namely: that speakers can choose the degree of delicacy they traverse in systems during instantiation, with greater delicacy corresponding to greater commitment. In fact, it is only by traversing the entire system to the most delicate features that lexical items can be specified.

[3] To be clear, abductive inferencing is not the best available mode of reasoning:

Abductive reasoning (also called abduction, abductive inference, or retroduction) is a form of logical inference that seeks the simplest and most likely conclusion from a set of observations. It was formulated and advanced by American philosopher Charles Sanders Peirce beginning in the last third of the 19th century. Abductive reasoning, unlike deductive reasoning yields a plausible conclusion but does not definitively verify it. Abductive conclusions do not eliminate uncertainty or doubt, which is expressed in retreat terms such as "best available" or "most likely".

In later years his view came to be: 

Abduction is guessing. It is "very little hampered" by rules of logic. Even a well-prepared mind's individual guesses are more frequently wrong than right.

Moreover, abductive reasoning is formally equivalent to a logical fallacy:

Abductive reasoning allows inferring a as an explanation of b. As a result of this inference, abduction allows the precondition a to be abduced from the consequence b. Deductive reasoning and abductive reasoning thus differ in the direction in which a rule like "a entails b" is used for inference. As such, abduction is formally equivalent to the logical fallacy of affirming the consequent (or Post hoc ergo propter hoc) because of multiple possible explanations for b.

[4] Trivially, the papers in this publication are from the 2007 conference at the University of Sydney titled 'Semiotic Margins'. Non-trivially, Martin's paper in this collection includes his classic misunderstanding of Saussure and Hjelmslev. Martin (2011: 245):

To be clear, Saussure's signifié and signifiant correspond to Hjelmslev's content and expression, and it is these that are the focus of linguistics. The relation between these two levels of symbolic abstraction is simply realisation. 

[5] With regard to the work by Martin and his colleagues on paralanguage, see:

• Buyer Beware!
• Martin's Model Of Paralanguage

[6] For a critique of some of the theoretical misunderstandings in these works, see the earlier post Yaegan Doran On The Metafunctions In Mathematical And Chemical Symbolism.

Bradley Smith On Music And Metafunction

Bradley Smith replied to Ed McDonald on Sysfling on 25/8/23 at 10:08:

in thinking about the semiosis/semiotics of music, it has always seemed to me that the same problem occurs as in intonation study: our scope for what we consider to be 'meaning'. In intonation study, much that I consider 'meaning' has been dismissed as emotional reflexes or some such, as though those reflexes aren't grounded in some communication of meaning! So, the same with music: what meanings are there in music, beyond those found in other semiotic modes? From there, in terms of metafunction, the next question I would ask is, can we see prosodies or periodicities in music, and how are they realised? I happen to have a fundamental faith in the theory of metafunction, distinguished via prosodic and periodic expression, I have found it essential to understanding how tone group and clause interact, together and with body language, for example, as in the recent Paralanguage book.

Blogger Comments:

[1] To be clear, to date, no-one has been able to devise system networks for the content plane of music, together with realisation statements specifying how the systems are realised structurally; nor has anyone been able to identify how content plane systems are realised by expression plane systems and their structures.

For the view that music is not a socio-semiotic system, but a social system expanded by socio-semiotic systems such as theory, notation and lyrics, see Making Sense Of Music.

[2] To be clear, prosodies and periodicities are the types of structure favoured by the interpersonal and textual metafunctions, so Smith's question is 'how are structures realised?'. In SFL Theory, function structures (e.g. Senser ^ Process) are realised by syntagms of classes of units at the lower rank (e.g. nominal group ^ verbal group). That is, the question presumes that music has not only a grammar, but a grammar with a rank scale of grammatical forms. For Halliday, on the other hand, language is the only semiotic system that has a stratified content plane with a grammar; see Halliday & Matthiessen (1999: 602ff).

[3] To be clear, here Smith gives priority to the view 'from below', structure, and suggests that structure types distinguish the metafunctions. SFL Theory, on the other hand, gives priority to view 'from above', so that it is the metafunctions that engender different types of structure.

[4] See 

• Modelling Paralanguage Using Systemic Functional Semiotics (Ngo et al. 2022)
• Buyer Beware!
• Martin's Model Of Paralanguage

David Rose Misunderstanding Metafunctionality And Axis (Inter Alia)

David Rose replied to Yaegan Doran and John Bateman on Sysfling on 25 August 2023 at 8:40:

This makes excellent sense. Metafunctionality as I tried to suggest, is a property of lg that its syntagmatic structuring has evolved to enable, together with ranks and strata. So it’s basic to lg but not to semiotic modes in general.

My original point was that axis is the basis of semiosis and that these are all dimensions in which axis proliferates in lg. John singled out metafunction as less basic than the others (for his own ends we now see).

What about ranks and strata? Lg has a stratified content plane and elaborate rank systems in grammar and phonology. Again metafunctionality is enabled by the evolved syntagmatic structuring of texts, clauses, groups, words and morphemes, together with tone groups, feet and syllables.

How much of these are basic to other modalities?

Blogger Comments:

[1] To be clear, here Rose is endorsing Doran's misunderstanding of the metafunctions in mathematical and chemical symbolism. For detailed reasons as to why it does not make excellent sense, see the previous post Yaegan Doran On The Metafunctions In Mathematical And Chemical Symbolism.

[2] To be clear, here Rose gives priority to the view 'from below', structure, and suggests that structure evolved to enable metafunctional meanings. This would mean, for example, that Theme^Rheme structure evolved to enable the textual metafunction. 

In SFL Theory, on the other hand, it is the other way round: priority is given to the view 'from above', system and meaning, and it is the different systems of metafunctional meaning that are realised as different types of structure. 

[3] It is not clear here, if Rose is suggesting that ranks and strata evolved to enable the metafunctions, or that ranks and strata were enabled by syntagmatic structuring. In any case, neither suggestion is tenable. Rank is a local dimension and the ordering principle of structure, and stratification is a global dimension whose organising principle is realisation. Metafunction is a distinct global dimension with a distinct ordering principle. Halliday & Matthiessen (2014: 20, 32):

[4] To be clear, the insights of SFL Theory suggest the opposite point of view: that no all-purpose semiotic system could evolve without the metafunctions. Halliday & Matthiessen (1999: 532-3):

These three "metafunctions" are interdependent; no one could be developed except in the context of the other two. When we talk of the clause as a mapping of these three dimensions of meaning into a single complex grammatical structure, we seem to imply that each somehow "exists" independently; but they do not. There are — or could be — semiotics that are monofunctional in this way; but only very partial ones, dedicated to specific tasks. A general, all-purpose semiotic system could not evolve except in the interplay of action and reflection, a mode of understanding and a mode of doing — with itself included within its operational domain. Such a semiotic system is called a language.

[5] To be clear, in SFL Theory, axis is a local dimension whose orders are the paradigmatic and syntagmatic. Since there are semiotic systems without syntagmatic chaining, such as protolanguage, axis cannot be the basis of semiosis. A more fundamental dimension is Saussure's distinction between the two components of the sign, signified and signifier, which SFL Theory models as the dimension of stratification: content and expression.

[6] To be clear, here Rose gives priority to the view 'from below', phonological form, and suggests that phonological form evolved to enable metafunctional meanings. This would mean, for example, that the foot evolved to enable the textual metafunction.

In SFL Theory, on the other hand, it is the other way round: priority is given to the view 'from above', content, and it is metafunctional content that is realised by phonological forms.

Yaegan Doran On The Metafunctions In Mathematical And Chemical Symbolism

Yaegan Doran wrote to Sysfling on 25 August 2023 at 7:09:

… we should be wary of metafunctions being considered a basic theoretical parameter for the description of semiotic resources in general.

A basic argument I put in a paper once was:

‘If we take metafunctionality to be the one of the broadest means by which these traditions [SFL/Social Semiotics] conceptualize the intrinsic functionality of semiotic resources, by simply assuming metafunctions across semiosis, we run the risk of homogenizing descriptions and making everything look like the first resource to be comprehensively described (i.e. English). That is, we risk watering down the specific functionality of each resource.’

Rather, following people working in language description at the time, I tried to tease these issues out for mathematical symbolism in the attached paper (and in more detail in my 2018 book Discourse of Physics), by attempting to ‘derive’ metafunctionality from axis (following Martin’s 2013 book on axis). Maths, I figured (with years of excellent help from many), would prove a useful point of contrast to language by virtue of regular common-sense comparisons that considered it to ‘be’ language/a language. But despite its comparability to language, I could find no evidence in the internal systemic/structural organisation of mathematical symbolism for an interpersonal component, and it was really pushing it to interpret an experiential component. But I tried to suggest that there was a major component comparable to the logical in language and a smaller independent one comparable to the textual. Nonetheless, I do think it much closer to language in this regard than most other resources, and so if even it struggles to fill out the full metafunctional complement, we should not assume it across the board (though metafunction is of course still a useful way in for analysis/application that is less concerned with descriptive/theoretical development – which I also tried to emphasise in the attached paper).

Along similar lines, Yu Zhigang followed this up by looking at the range of symbolism used in chemistry and concluded similarly – that the logical dominated, that there was little evidence for an interpersonal, that there was an even smaller textual component than maths, but there was in fact something comparable to an experiential. But he went a little further, and noted that some structuring principles that we often associate with interpersonal meaning in language – i.e. prosodic structure – occurred in various chemical symbolisms, but seemed to have become ‘ideationalised’ in that they were used to realise meaning related to field, not tenor.

Blogger Comments:

[1] To be clear, this is not just an argument against the metafunctional dimension of the theory, but against any dimension of the theory, and so against the theory itself, since in this view, using a theory that was devised for English, SFL, runs 'the risk of homogenising descriptions and making everything look like the first resource to be comprehensively described'. The same argument could be made against generalising Newton's theory of gravitation from falling apples to orbiting heavenly bodies, since it runs 'the risk of homogenising descriptions and making everything look like the first resource to be comprehensively described'. 

[2] To be clear, the metafunctions cannot be derived from axis, even if axis is properly understood. This is not least because they are independent dimensions, with different scopes, and organised according to different principles. Halliday & Matthiessen (2014: 20, 32):

That is, the local distinction between the paradigmatic and syntagmatic orders of axis entails nothing about the global distinctions between construing experience as meaning, enacting intersubjective relations as meaning, and creating text from the two.

For some of the theoretical misunderstandings of axis in Martin (2013), see the review here.

[3] To be clear, from the perspective of SFL Theory, mathematics is a contextual semiotic that is realised in language (e.g. algebraic equations) and a pictorial semiotic (e.g. geometric diagrams). The proof that equations, for example, are language is that, unlike non-language, they can be read aloud. This is because what is projected by saying, locutions, constitutes the lexicogrammatical stratum, and language is the only semiotic system with a stratified content plane. This is why bi-stratal systems, like pictorial semiotics, cannot be read aloud.

[4] To be clear, this is the logical fallacy known as the Argument from ignorance:

Argument from ignorance (from Latin: argumentum ad ignorantiam), also known as appeal to ignorance (in which ignorance represents "a lack of contrary evidence"), is a fallacy in informal logic. It asserts that a proposition is true because it has not yet been proven false or a proposition is false because it has not yet been proven true. This represents a type of false dichotomy in that it excludes the possibility that there may have been an insufficient investigation to prove that the proposition is either true or false.

Here Doran has misconstrued absence of evidence as evidence of absence. The evidence for the interpersonal metafunction in mathematical symbolism is manifold. For example, mathematical symbolism is language, so it enacts interpersonal meaning. As language, the equation E = mc² is a proposition realised by a declarative clause that can argued as valid or not whether it is expressed this way or as energy equals mass times the speed of light squared. If there were no interpersonal meaning in mathematical symbolism, then mathematicians would not be able to argue for or against their validity.

[5] To be clear, mathematical symbolism is deployed in many sciences precisely because it construes experience as meaning. As language, a mathematical equation is a figure of being realised by an intensive identifying relational clause, and the solving of an equation is the iterated elaboration of the clause until one or more unknown Tokens is decoded by reference to a Value.

[6] To be clear, the logical metafunction is concerned with the logico-semantic relations of expansion and projection, so it requires experiential meanings to relate logically to one another. Every equation is an intensive identifying clause that construes an elaborating logical relation between a Token and a Value, and the solving of an equation entails a clause complex in which all clauses are logically related through elaboration. The logical relation of extension, for example, is manifested in processes of addition and subtraction.

[7] To be clear, none of the metafunctions is independent of the others. Halliday & Matthiessen (1999: 532-3):

These three "metafunctions" are interdependent; no one could be developed except in the context of the other two. When we talk of the clause as a mapping of these three dimensions of meaning into a single complex grammatical structure, we seem to imply that each somehow "exists" independently; but they do not. There are — or could be — semiotics that are monofunctional in this way; but only very partial ones, dedicated to specific tasks. A general, all-purpose semiotic system could not evolve except in the interplay of action and reflection, a mode of understanding and a mode of doing — with itself included within its operational domain. Such a semiotic system is called a language.

More specifically, the textual metafunction requires the other metafunctions, since its function is to operationalise them. Halliday & Matthiessen (1999: 7-8, 398):

The textual metafunction is an enabling one; it is concerned with organising ideational and interpersonal meaning as discourse — as meaning that is contextualised and shared. … 

The function of the textual metafunction is thus an enabling one with respect to the rest; it takes over the semiotic resources brought into being by the other two metafunctions and as it were operationalises them…

Every mathematical equation is organised in terms of Theme and information; see [4] above.

[8] To be clear, given the explanatory power of the metafunctions, it would be more prudent to assume that they are a dimension of a semiotic system until it is proved otherwise. As explained above in [4], Doran mistook his inability to identify the metafunctions in mathematical symbolism for evidence that they are not a dimension of its systems. 

[9] To be clear, Yu Zhigang is Doran's collaborator, rather than a scholar working independently of Doran. More importantly, what is said above about mathematical symbolism also applies to chemical symbolism. For example:

∞

See also Yaegan Doran Failing To Recognise The Interpersonal Metafunction In Mathematics.