Confusing Negative Addition With Negative Polarity

Martin (1992: 205-6):

External additive relations, like external comparative ones, form a small resource compared with temporals or consequentials.  The basic distinction is between addition and alternation.  With addition, there is a positive/negative system opposing and to nor, in spite of morphological appearances which make it look like a negative realisation of or (Our Corgi didn't win nor did the Dachshund means that 'the Corgi didn't win and that the Dachshund didn't either').  All additive relations have correlative paratactic realisations [both…and, neither…nor, either…or].

[4:105] [addition:positive]
             a.  Our Dachshund performed well.  And she looked splendid on the day.
             b.  Besides performing well our Dachshund looked splendid on the day.
             c.  – 

[4:106] [addition:negative]
             a.  She didn't move at all well in the ring.  Nor did she stand still when tabled.
             b.  Alongside not moving well in the ring, she didn't stand still when tabled.
             c.  – 

[4:107] [alternation]
             a.  You could go down to Melbourne.  Or you could go in the Easter Show.
             b.  If you don't go down to Melbourne, you could go in the Easter Show.
             c.  –

Blogger Comments:

[1] In SFL theory, addition, alternation and variation are the three main categories of extension relations between clauses.  In contrast, in Martin's model:

1. both addition and alternation are subsumed under the category additive, 
2. variation subtypes are construed as subcategories of adversative, and 
3. adversative — a subcategory of addition in SFL — is construed as a subcategory of comparison — a subcategory of the enhancement category manner in SFL.

Halliday & Matthiessen			Martin
extension	alternation		alternation	additive
	addition	additive: positive	addition: positive
		additive: negative	addition: negative
		adversative*
	variation	replacive	replacement	adversative	comparison
		subtractive	exception

[2] The logical relation here is positive, not negative addition — hypotactic negative addition does not occur in English.  The negation here is realised as polarity in verbal groups, not as a logical relation between clauses.

Using Clause Simplexes To Theorise Conjunctive Relations

Martin (1992: 204):

There are two variations on [similarity], one conditional and one temporal, where apparently hypotactically dependent clauses cannot appear initially in the clause complex:

SIMILARITY (conditional)

[4:99]  Ben looked

            as if he'd just won the Best of Show. 

SIMILARITY (temporal)

[4:100] Ben looked

            like when he won Best of Show.

Blogger Comments:

[1] Neither of these is a clause complex; each is an attributive clause.  If Ben looked were a clause, it would be one without an Attribute.

In the first, an embedded clause serves as Attribute:

Ben	looked	[[as if he'd just won the Best of Show]]
Carrier	Process: attributive	Attribute/Manner: comparison

In the second, an embedded clause complex (with Mood ellipsis) serves as Attribute:

Ben	looked	[[like (he did) || when he won Best of Show]]
Carrier	Process: attributive	Attribute /Manner: comparison

[2] Both Attributes construe comparison (as if, like).

Misconstruing Subtractive And Replacive Variation (Extension) As Subtypes Of Comparison (Enhancement)

Martin (1992: 204):

In addition to the kind of contrast illustrated in [4:95], there are replacement and exception categories to consider.  With replacives (e.g. instead of, in place of) the blurry line between contrast and alternation discussed in 4.2.2 above is hardest to draw:

EXCEPTION

[4:97]  We had a good day

            except for losing the Best of Show award. 

REPLACEMENT

[4:98]  We won Best of Show

            instead of missing out as usual.

Blogger Comments:

[1] In SFL theory, the logical relation in [4:97] is extension: variation: subtractive ('except').  Its meaning is X but not all X.

[2] In SFL theory, the logical relation in [4:98] is extension: variation: replacive ('instead').  Its meaning is not X but Y.

The basic opposition that is proposed for contrast within 'external comparison' (enhancement) is thus actually between subtractive variation (extension) and replacive variation (extension).

Note that the entry condition to this opposition, 'contrast', corresponds in SFL theory to a subtype of addition (adversative), which opposes variation and alternation within extension.

[3] The logical meaning of alternation is X or Y; the logical meaning of 'contrast' (adversative addition) is X and conversely Y.

Misconstruing Adversative Addition (Extension) As Manner: Comparison (Enhancement)

Martin (1992: 203):

The external comparison network is a small one; the basic opposition is between [contrast] and [similarity]:

CONTRAST

[4:95]  Whereas usually we win

            this time we lost. 

SIMILARITY

[4:96]  As usually happens

            we won.

Blogger Comments:

[1] This confuses the notion of contrast (A in contrast to B) with the notion of dissimilarity (A is unlike B).

[2] In SFL theory, the logical relation in [4:95] is extension: addition: adversative ('but').  Its meaning is X and conversely Y: usually we win and conversely (but) this time we lost.  (The logical relation in [4:96] is manner: comparison — N is like M).

The basic opposition that is proposed for 'external comparison' (enhancement) is thus, in SFL terms, between adversative addition (extension) and manner: comparison (enhancement).

The Argument For Comparison As A Major Logical Category

Martin (1992: 203):

Taking into account these textual patterns and the fact that a similar/different opposition is basic to all other discourse semantic systems, comparison will be set up as a major category of conjunctive relation here.  The congruence of the like/unlike opposition across discourse systems is shown below. 

Table 4.11 Contrast and similarity across discourse semantic systems
	“LIKE”	“UNLIKE”
CONJUNCTION	similarity	contrast
IDENTIFICATION	semblance	difference
IDEATION	synonymy	antonymy
NEGOTIATION (Halliday 1985: 69)	accept undertake acknowledge answer	reject refuse contradict disclaim

Blogger Comment:

[1] As demonstrated in previous critiques, the analysis of 'these textual patterns' is compromised by the logical error of confusing different/unlike with contrastive/adversative.

[2] The opposition of similar/different and of like/unlike demonstrates the semantic distinction between adversative (opposition) and dissimilar (different/unlike) — dissimilar is one of the elements in opposition.

[3] Difference is the basis of semiosis itself, not just discourse semantic systems.  Similarity is the basis of agnation, including metaphorical agnation.

Conclusion:  The argument for setting up comparison as a major category of conjunctive relations rests on two misunderstandings:

1. confusing enhancement: manner: comparison with extension: addition: adversative, and
2. presenting a fundamental feature of all semiosis as specific to discourse semantic systems.

Confusing Dissimilar (Enhancement) With Adversative (Extension)

Martin (1992: 202-3):

Consider now a modified version of this part of the interview:

[4:94] a. With the bigger breeds of dog, they're stood on the ground.

          b. With the smaller breeds such as Dachshunds, they're stood on the table.

          c. Likewise breeds with similar stature like the Corgis, all the Toy-breeds, and this type of thing are tabled,

          d. so as the judge can handle them more easily.

This text focusses on similarity as well as difference.  Lexical cohesion (the co-hyponymy of Dachshunds with Corgis and Toy-breeds and the repetition of table (table, tabled), comparative reference (similar stature) and an explicit conjunction (likewise) make the comparison clear.

Halliday and Hasan (1976) treat the contrastive relation in [4:94: a–b] as closely related to concession (under the general adversative category) and the similarity in [4:94: b–c] as a kind of additive.  Halliday (1985) on the other hand treats contrast as related to additive relations under the heading extension and similarity as closely related to temporals and causals under the heading enhancement. …

These interpretations fail to bring out the very similar ways in which IDENTIFICATION and IDEATION participate in the realisation of both contrast and similarity (as exemplified in [4:84sic] above…

Blogger Comment:

This continues the category error of conflating different (dissimilar) with contrastive (adversative).  The logical meaning of dissimilar is N is not like M, whereas the logical meaning of adversative is X and conversely Y.  In SFL theory, the logical relation of dissimilar is a subtype of manner, within enhancement, whereas the logical relation of adversative is a subtype of addition, within extension.  That is, dissimilar and adversative belong to different major categories within expansion.

Confusing Enhancement (Manner: Comparison) With Extension (Adversative Addition)

Martin (1992: 202):

Within the framework of systemic functional approaches to discourse, comparative relations are the most controversial category to be developed here; … they are not treated as a major category of logical relation by either Halliday and Hasan [1976] or Halliday [1985].  To explore this question, it will be useful to return to text [4:2].

[4:2] p. With the bigger breeds of dog, they're stood on the ground,
        q. because it's easier for the judge to handle them.
       r. With the smaller breeds of dog, such as Corgis, all the Toy-breeds, Dachshunds and this type of thing we — as our turn comes,
        s. we stand our dog on the table.

As noted in Section 4.1 above, there is a clear relationship of contrast between p–q and r–s.  This is coded through the comparative reference (bigger vs smaller) and the lexical cohesion (most clearly ground vs table); and the contrast is highlighted by the marked Themes in p and r.  As noted in 4.2.4 above the contrast could have been made explicit in [4:2] with a conjunction such as but.

Blogger Comment:

[1] This confuses two distinct logical relations, whose meanings are N is like M and X and conversely Y.  In SFL theory, the former is classified as enhancement: manner: comparison, and the latter as extension: additive: adversative ('but').

[2] This fundamental confusion is one factor that explains why this approach is "controversial", not least because it undermines the logical validity of what follows.

Misconstruing Positive Condition As Concessive

Martin (1992: 200):

Concessive conditionals can be further subclassified according to the universality of the contingency denied.  Single, alternative and universal concessive contingency is illustrated below:

CONDITIONAL/CONCESSIVE: SINGLE

[4:81]  Even if we get that judge again

            we'll win. 

CONDITIONAL/CONCESSIVE: ALTERNATIVE

[4:81]  Whether we get that judge again or not

            we'll win. 

CONDITIONAL/CONCESSIVE: UNIVERSAL

[4:81]  Whichever judge we get

            we'll win.

Blogger Comments:

[1] Neither the 'alternative' nor the 'universal' agnate involves the logical relation of condition: concessive — if P then contrary to expectation Q.  Both involve the logical relation of positive condition — if P then Q.  Winning is not contrary to expectation under either condition.

[2] This confuses enhancement (condition) with extension (alternation).  Here the logical relation of extension: alternation — X or Y — is realised in the clause nexus forming the condition (whether … or not).

[3] Therefore, the difference between the three conditionals is not "according to the universality of the contingency denied".

Misconstruing Adversative Extension As 'Concessive Purpose'

Martin (1992: 199):

The hypotactic proportionalities can be exemplified as follows:

NON-CONCESSIVE : CONCESSIVE :: 

(MANNER)

Ben improved his time by training hard :

Ben didn't improve his time even by training hard :: 

(CONSEQUENCE)

Ben improved his time because he trained hard :

Ben didn't improve his time even though he trained hard :: 

(CONDITION)

Ben will improve his time if he trains hard :

Ben won't improve his time even if he trains hard :: 

(PURPOSE)

Ben can train hard in order to improve his time :

Ben can train hard without improving his time

Blogger Comments:

In SFL theory, the logical meaning of condition: concessive is if P then contrary to expectation Q.  It may be ordered concession^consequence or consequence^concession.

[1] The logical relation here does not involve manner: means — N is by means of M.  A manner agnate would be Ben didn't improve his time by means of training hard.  The logical relation here is simply condition: concessive as in Ben didn't improve his time even though he trained hard.

[2] The logical relation here does not involve cause: reason (Martin's 'consequence') — because P so result Q.  A reason agnate would be because he trained hard Ben didn't improve his time.  The logical relation here is simply condition: concessive.

[3] The logical relation here does not involve cause: purpose — because intention P so action Q.  The logical relation here is not even a type of enhancement, but the type of extension termed addition: adversative — X and conversely Y — as in the paratactic agnate Ben can train hard and not improve his time.

Misconstruing Concession As Manner

Martin (1992: 198-9):

It remains to consider concessive relations, which will be taken here as crossclassifying manner, consequence, condition and purpose rather than as a fifth consequential category.  The reason for this is that concessives function as a kind of "anti-modulation", cancelling the potentiality or obligation which would otherwise enable or determine the consequential relation between events.

The opposition is illustrated for manner relations in [4:79] and [4:80]; in [4:79] entering the right shows was enough to win, but in [4:80] the enablement is cancelled — entering the right shows was not sufficient to win.

MANNER

[4:79]  By entering the right shows,

            we won. 

MANNER/CONCESSION

[4:80]  Even by entering the right shows,

            we didn't win.

 Blogger Comments:

[1] In SFL theory, the logical relation of concession is a type of condition, with the logical meaning of if P then contrary to expectation Q.  As such, it does not cross-classify the other enhancement relations, as demonstrated more fully in tomorrow's posting.

[2] This confuses the logical relation of concessive condition with the interpersonal system of modality.

[3] This confuses interpersonal modality ("enablement") with sufficient conditions — here contrasted with insufficient conditions rather than necessary conditions.

[4] The logical relation here does not involve manner: means — N is by means of M.  A manner agnate would be we didn't win by means of entering the right shows.  The logical relation here is simply condition: concessive as in we didn't win even though we entered the right shows.

Misconstruing A Dependent Clause As A Postmodifier In An Adverbial Group

Martin (1992: 198):

Consequence is also commonly realised referentially, through esphoric comparison as in [4:78]:

[4:78]  We prepared so well [[that we won]].

Blogger Comments:

[1] The clause that we won is not embedded.  Instead, it is a ranking dependent clause in a nexus where the logical relation is cause: result; because P (we prepared so well), so result Q (we won).

we	prepared	so well	that	we	won
a			x b (cause: result)
Actor	Process: material	Manner: quality		Actor	Process: material

[2] The wording so well that we won does not construe comparison (for which, see Halliday & Matthiesen 2004: 357-8).  Adverbial groups with an embedded clause as a Post-modifier which do construe comparison are those like the following:

better	[[than we needed]]
Head	Post-modifier

as in:

we	prepared	better [[than we needed]]
Actor	Process: material	Manner: quality

[3] Consequence is not realised referentially in the instance intended to demonstrate same.  Instead, cause: result is realised as a logical relation between two ranking clauses.

Misconstruing Manner As A Cause-Effect Relation

Martin (1992: 198):

Both manner (whereby, thereby) and consequence (/ ^ so that /) have realisations which apparently make the Effect dependent, but which do not allow the beta-clause to be realised first in the clause complex:

MANNER

[4:76]  We prepared well

            thereby managing to win. 

CONSEQUENCE

[4:77]  We prepared well

            / ^ so that / in the end we won.

Blogger Comments:

[1] The logical relation of manner does not involve an effect of a cause.  The conjunction thereby, however can realise both manner and cause: reason/result.

[2] In both SFL theory and formal logic, the term 'consequence' is used with condition (if P then Q), not with cause: reason/result (because P so result Q). 

Misconstruing Reason And Result As Purpose

Martin (1992: 197):

Degrees of probability are also possible with purposives, but there are only two values, high/median and low:

PURPOSIVE: DESIRE: MODALITY POSSIBLE

[4:74]  We'll enter

            in case we win. 

PURPOSIVE: DESIRE: MODALITY CERTAIN/PROBABLE

[4:75]  We'll enter

            so that we'll win.

Blogger Comments:

[1] This confuses the logico-semantic relation of cause with interpersonal modality values.

[2] This confuses the logical meaning of purpose, because intention Q so action P with desire.

[3] Although this is intended to demonstrate a difference in 'purposives', the logico-semantic relation of cause: purpose is not a feature of either clause nexus.

The logico-semantic relation in the first 'purposive' clause nexus is cause: reason.  The logical meaning is Q because P.  Winning is construed as the reason for entering, not the purpose of entering: we'll enter because we might win. (Hence the interpretation as low probability).

The logico-semantic relation in the second 'purposive' clause nexus is cause: result.  The logical meaning is because P so result Q.  Winning is construed as the result of entering, not the purpose of entering: we'll enter with the result that we'll win.  (Hence the interpretation as high/median probability).

[See Halliday and Matthiessen (1999: 245) for why tying the relator to the effect in cause: result sequences typically implies intentionality.]

The upshot is that the two 'purposive' clause nexuses differ in terms of type of cause — neither of which is purpose.

Confusing Condition With Probability

Martin (1992: 197):

Inclusive conditionals can be further divided into factual and counterfactual; with counterfactuals the beta-clause [omitted verb] an additional secondary [past] tense.

COUNTERFACTUAL

[4:70]  If we had prepared well,

we would have won.

Factual conditionals choose among high, median and low modalisation, according to the probability of the Cause taking place:

FACTUAL: MODALITY POSSIBLE

[4:71]  If we enter

           we'll win. 

FACTUAL: MODALITY PROBABLE

[4:72]  Provided we enter

           we'll win. 

FACTUAL: MODALITY CERTAIN

[4:73]  As long as we enter

           we'll win.

Blogger Comments:

[1] This confuses the logico-semantic relation of condition with the truth values of the meanings realised by the clauses conditionally related in a nexus.

(In formal logic, a counterfactual conditional — in contradistinction to a material conditional — is a subjunctive conditional containing an if-clause which is contrary to fact.)

[2] This confuses the logico-semantic relation of condition with interpersonal modality values.  Within that error, it wrongly claims that the beta-clauses of three clause nexuses differ in probability values.  In terms of logical meaning, each nexus is simply if P then Q.

If the nexuses had differed in modalisation values, the 'factual conditionals' would've looked something like the following, with the modalisation values enacted in the alpha-clauses:

If/Provided/As long as we enter, we'll possibly win

If/Provided/As long as we enter, we'll probably win

If/Provided/As long as we enter, we'll certainly win

 [3] Because the logico-semantic relation is condition, rather than cause: reason, there is no cause-effect relation construed in these clause nexuses.

If the logico-semantic relation had been reason, there would've been a cause-effect relation construed in the clause nexuses.

Misconstruing Condition And Purpose

Martin (1992: 196-7):

But with conditionals, the Effect may or may not be desirable; if and unless do not code it as one or the other:

[4:66] If you go that way,

           you'll get there by six/you'll get lost. 

[4:67] Unless you go that you [way]

           you'll get there by six/you'll get lost.

Purposives on the other hand always code the Effect as desired or feared.  In [4:68] Ben will be read as wanting to get lost, and as not wanting to get there by six in [4:69].

[4:68] Ben went that way

           so that he'd get lost. 

[4:69] Ben went that way

           for fear of getting there by six.

In other words, fear and desire are implicit in [4:66] and [4:67], but can be read in where the context indicates they are needed to make sense of the conditional meanings made explicit.  Fear and desire on the other hand are grammaticalised in [4:68] and [4:69].

Blogger Comments:

[1] Clauses related by the logico-semantic relations of condition and purpose do not construe a cause-effect relation.  Condition means if P then Q, purpose means because intention Q so action P.  It is reason/result — because P so result Q — that construes a cause-effect relation.

[2] Because the logical meaning of purpose involves intention, not desire and fear, desire and fear are not grammaticalised in clauses logically related by cause: purpose.

[3] Because the logical meaning of condition does not involve desire and fear— cf if it has three sides, it is a triangle — desire and fear are irrelevant to the discussion of clauses logically related by condition.

Misconstruing Negative Vs Positive Purpose

Martin (1992: 195-6):

Contingent relations make a distinction between conjunctions incorporating negative polarity (unless and lest) and those which don't.  The relevant proportionalities and relevant paradigm are as follows:

UNLESS : IF…NOT ::

LEST : SO THAT…WILL + NOT 

unless Ben plays you'll lose :

if Ben doesn't play you'll lose :: 

Ben'll play lest you lose :

Ben will play so that you won't lose

But the opposition between "positive" and "negative" values has a different meaning in the context of conditional relations from that in purposives. … With purposives, the opposition is between [desire] and [fear].  So that encodes a inclination to achieve the Effect, lest an inclination to avoid it:

DESIRE

[4:62]  He went that way

            so that he'd get there by six.

FEAR

[4:63]  He went that way

            lest he lose his way. (less archaically: for fear of losing his way.)

Blogger Comments:

[1] Because it is a category error to construe the logical meaning of purpose, because intention P, so action Q as 'desire (for Effect)', it is also a category error to construe the negative agnate as 'fear (of Effect)'.

[2] This confuses the logical and interpersonal metafunctions.  Just as the logical relation of purpose is distinct from the interpersonal feature inclination, so too is the negative agnate.

[3] Compounding the previous confusion, the terms 'achieve' and 'avoid' distinguish the positive and negative poles of inclination in terms of conation, instead of dis/inclination.

[4] The logical meaning of purpose does not include the relation of cause and effect.  In the 'purpose' nexus she cast a spell to ward off evil spirits, warding off evil spirits is the purpose, but not the effect of casting a spell.  Cause and effect appear in complexes where clauses are related by cause: reason (or cause: result).

Misconstruing Negative Vs Positive Condition

Martin (1992: 195-6):

Contingent relations make a distinction between conjunctions incorporating negative polarity (unless and lest) and those which don't.  The relevant proportionalities and relevant paradigm are as follows:

UNLESS : IF…NOT ::

LEST : SO THAT…WILL + NOT 

unless Ben plays you'll lose :

if Ben doesn't play you'll lose :: 

Ben'll play lest you lose :

Ben will play so that you won't lose

But the opposition between "positive" and "negative" values has a different meaning in the context of conditional relations from that in purposives.  With conditionals, the opposition is between exclusion and inclusion (or non-exclusive) to be precise).  Unless means 'if and only if not';  if on the hand does not preclude the possibility of additional modalised Causes:

EXCLUSIVE

[4:60]  Unless you go that way ['as long as you don't]

            you'll be there by six.

            (It's the only way you can go wrong.) 

INCLUSIVE

[4:61]  If you go that way

            you'll be there by six.

            (But you could also go the back way.)

Blogger Comment:

In SFL theory, as in formal logic, the logical meaning of positive condition is if P then Q and the logical meaning of negative condition is simply if not P then Q — not "if and only if not":

• If Ben doesn't play, you'll lose. 
• If you don't go that way, you'll be there by six.

The opposition of 'exclusive vs non-exclusive' is irrelevant to the logical opposition of positive and negative condition.

See also if and only if.

Misrepresenting Reason As Purpose

Martin (1992: 194):

The distinction between condition and purpose has to do with modulation again. Purposives contain an additional modulation of inclination, associated with the Effect (the Effect is desired). This can be brought out by comparing hypotactic and paratactic realisations of the purposive relation; these proportionalities also reveal another peculiarity of purposive relations, namely that desire for Effect commences before the Cause — wanting to win, get there on time and give the opposition a chance are the motivation for, not the results of, training hard, driving fast and skating slowly in the examples below (with all other consequential relations the Cause is temporally anterior):

we trained hard so that we'd win :

we wanted to win, and so we trained hard :: 

we drove fast in order to get there on time :

we were keen to get there on time and so we drove fast :: 

we skated slowly to give them a chance :

we were willing to give them a chance and so we skated slowly

Blogger Comments:

This is the second of two critiques of this extract.

[1] Contrary to the claim, none of the paratactic clause complexes involves the logical relation of cause: purpose.  In all three, the relation is cause: reason (see Halliday & Matthiessen 2004: 411).  This can be demonstrated by substituting and therefore for and so:

we wanted to win, and therefore we trained hard
we were keen to get there on time and therefore we drove fast
we were willing to give them a chance and therefore we skated slowly

The proportionality is thus hypotactic purpose : paratactic reason.

The paratactic clauses, therefore, unlike the hypotactic clauses, do construe a cause^effect relation.

[2] In each of the three paratactic clause complexes (of cause: reason), 'desire' is a feature of the clause construing the cause.  That is, the "desire for Effect" does not "commence before" the Cause.