Confusing Collocation With Transitivity

Martin (1992: 309):

In previous approaches to lexical cohesion, nuclear relations have been handled under the heading collocation.  An attempt will be made to unpack these relations here in order to identify more precisely the semantic relations involved.  What this amounts to is a foray into the discourse semantics of experiential grammar, which is in itself a daunting task.  It is however an essential one, since the lexical relations under consideration here cannot be explained simply by appealing to grammatical structure.  The relation between serve and ace for example is not limited to the Process Medium structure itself; the elements configured may be in different clauses — Ben serves… That's his fifth ace of the match.  And the configuration may be realised metaphorically — Ben's serve produced very few aces today.

Blogger Comments:

[1] In SFL theory, collocation is a type of lexical cohesion, and so a non-structural resource of the textual metafunction.  Clause nuclearity, in contrast, is grammatical: an aspect of the ergative model of transitivity, a structural resource of the experiential metafunction.

[2] An SFL model of ideational semantics that is theoretically consistent with SFL grammatics is expounded in Construing Experience Through Meaning (Halliday & Matthiessen 1999).

[3] In SFL theory, the semantic relations involved in collocation 'cannot be explained simply by appealing to grammatical structure'.  This is because the cohesive resource of collocation is lexical and non-structural.  On the other hand, instances of collocation can be sub-classified in terms the type of logico-semantic relation between the collocated lexical items.

[4] The relation between serve and ace is not one of Process and Medium when they appear as elements of the same clause, as in she served an ace.  An ace is not the Medium through which the serving Process unfolds; an ace is the Range of the serving Process. The Medium through which the serving Process unfolds is the server.

[5] When the elements serves and ace occur in different clauses, they are not configured, and so they are not related as Process + Medium.  The functions Process and Medium do not figure independently of clause structure.  In the example given, ace is part of the Range of a different Process (decoding identifying, rather than material).

Ben	serves
Medium	Process

that	’s	his fifth ace of the match
Medium	Process	Range

[6] In the metaphorical construal, serve functions as Agent, and aces functions as Medium:

Ben's serve	produced	very few aces	today
Agent	Process	Medium	Location

In a more congruent construal, served functions as Process, and aces functions as Range:

Ben	served	very few aces	today
Medium	Process	Range	Location

Misconstruing Ranges As Mediums

Martin (1992: 309):

Nuclear relations reflect the ways in which actions, people, places, things and qualities configure as activities in activity sequences.  The following Process Medium structures from the field of tennis illustrate the kind of relations involved.  In this field, players serve aces and volley winners, but they do not *serve lobs or *volley double-faults.

serve + ace

smash + overhead

put away + volley

hit + winner

net + passing shot

intercept + volley

drop + shot

lob + return

volley + winner 

Blogger Comments:

[1] This continues the misconstrual of field as the semantic description of social activities.  That is, context — the culture as semiotic system — is misconstrued as the language that describes social activities.  See previous critique here.

[2] These structures are all Process + Range, not Process + Medium.  (The tennis player is the Medium of each Process.)  In each case, the participant refines the process by naming a particular variety of it.  On Range, Halliday & Matthiessen (2004: 295) write:

There may be in each type of clause one element which is not so much an entity participating in the process as a refinement of the process itself.  This may be the name of a particular variety of the process…

[3] Ordinarily, this is construed nominally, rather than clausally, as Classifier + Thing.

An Unwarranted Claim About the System Of Identification

Martin (1992: 309):

This interaction between IDENTIFICATION and IDEATION provides a criterion for determining when relations between lexical items are cohesive.  They can be treated as such where one item codes sufficient experiential meaning to be presumed by another, either directly, or indirectly through bridging.

Blogger Comments:

[1] This was not demonstrated to be the case — even if the (previously discussed) confusion between reference and lexical cohesion in the discourse system of IDENTIFICATION is ignored. The examples used to argue the case were all clear cases of lexical cohesion: hyponymy (flowers-roses, insects-mosquitos) and meronymy (stadium-track, chimney-cottage). The system of IDENTIFICATION provided no criterion — it was irrelevant — for determining whether there was a lexically cohesive relation.

[2] In SFL theory, relations between lexical items are cohesive when items, as bundles of lexical features, are related by either repetition, synonymy, hyponymy, meronymy or collocation.

Misrepresenting Meronymy As Bridging Meronymy

Martin (1992: 308):

Bridging is also possible with parts and wholes, working in either direction:

BRIDGING: PART FROM WHOLE
[5:23] Ben walked into the stadium, to look at the track. 

BRIDGING: WHOLE FROM PART
[5:24] Ben saw a chimney. He was close to the cottage now.

Blogger Comments:

These are just clear cases of meronymy (lexical cohesion).  The underlined grammatical reference item the is irrelevant to the cohesiveness of the lexical relation.

Misrepresenting Hyponymy As Bridging Hyponymy

Martin (1992: 308):

Hyponyms do not directly presume their superordinates, but they may presume them indirectly through bridging. … In the following examples hyponyms are bridged from their superordinates:

BRIDGING: HYPONYM FROM HYPERONYM

[5:23]  Ben admired his flowers, especially the roses.  

BRIDGING: HYPONYM FROM HYPERONYM

[5:24]  Ben hated the insects, especially the mosquitos.

Blogger Comment:

These are just clear cases of hyponymy.  The underlined grammatical reference item the is irrelevant to the cohesiveness of the lexical relation.

Confusing Lexical Cohesion And Reference

Martin (1992: 308):

One way of resolving the question of how taxonomic distance affects cohesion is to make use of IDENTIFICATION and test items to see whether they are experientially close enough for one to have been presumed by the other.  Along these lines, the flower straightforwardly presumes a rose, as the insect presumes a mosquito.  Hyponyms do not directly presume their superordinates, but they may presume them indirectly through bridging.  The reason for this difference is that superordinates contain less information than their hyponyms, while hyponyms encode all of the experiential meaning of their superordinates.  This asymmetry is reflected in definitions:

rose  a beautiful & usually fragrant flower usually of red or yellow or white

Blogger Comments:

[1] This continues the confusion of grammatical reference and lexical cohesion, as previously explained here.

[2] Whether the reference item the in the flower presumes that the identity ('which flower?') is recoverable in a rose depends crucially on the co-text.  On the other hand, what is straightforward is the hyponymic relation between flower and rose.

Likewise, whether the reference item the in the insect presumes that the identity ('which insect?') is recoverable in a mosquito depends crucially on the co-text.  On the other hand, what is straightforward is the hyponymic relation between insect and mosquito.

[3] In SFL theory, lexically cohesive hyponymic relations do not involve presumption.

[4] The relation between superordinates and hyponyms is delicacy.  Superordinates are more general (broader) categories, hyponyms are more specific (narrower) categories.

[5] In terms of the SFL theoretical architecture, hyponymy combines (logical) elaboration with (experiential) attribution.  Accordingly, there is no encoding in hyponymy, because encoding is a direction of coding in identifying relations.  The experiential relation between hyponym and superordinate is class membership, not identity.

[6] Any informational asymmetry in the attributive relation between hyponyms and superordinates (Carriers and Attributes) is not "reflected" in the identifying relation between words and meanings (Tokens and Values).

Misconstruing Synonymy + Collocation As Relational Meronymy

Martin (1992: 307):

The second type [of collective item] cannot function as a Pre-Numerative: *a cattle of cows and calves or *a furniture of chairs and sofas are nonsense.  These items do nevertheless assemble a potentially heterogenous [sic] collection of individuals into a single group.  They function "cohesively" in examples like Ben put the cutlery, glasses and plates into the sink and washed up the dishes.  This set includes cattle, clergy, furniture, dishes, cutlery, beverages, parts of speech, shrubbery, collection, constellation, galaxy, solar system, compilation, assembly, company, conference, stable, troupe, orchestra, band, conglomeration, portfolio.

Blogger Comments:

[1] At least 12 of the 22 examples can function as Numeratives; these are all extended Numeratives of the type aggregate (collective measure).

[2] The lexical cohesion between dishes and plates is synonymy, and the lexical cohesion between dishes and each of cutlery and glasses is collocation.  Martin would appear to think that 'dishes' refers to an assembly of cutlery, glasses and plates.

[3] This "collective item" is actually a nominal group that includes an extended Numerative: facet (partitive + type) — in contradistinction to aggregates, which are collective + measure.

Misconstruing Ellipsis Of Repetition As Relational Meronymy [3]

Martin (1992: 306):

One set [of collectives] functions in Pre-Numerative position, as with the measure meronymy introduced above.  But here there function is to group together individuals rather than separate them out by parts (thus a member of the flock contrasts with a flock of geese).  Seen from the perspective of English's number system, their function is to reconstruct count nouns as a mass.  Like other items functioning in Pre-Numerative structures, they can be used cohesively between clauses: Did you see the ships? — I saw the whole squadron.

Blogger Comments:

[1] In SFL theory, this nominal group function is termed 'extended Numerative: aggregate', which is the cross-classification of 'collective' and 'measure'.  See Halliday & Matthiessen (2004: 333).

[2] In SFL theory, the contrast here is between portion (partitive measure) and aggregate (collective measure).  In partitives, Head < Thing (e.g. member < flock), whereas in collectives, Head > Thing (e.g. flock > geese).

a	member	of	the	flock
Numerative: portion			Deictic	Thing

a	flock	of	geese
Numerative: aggregate			Thing

[3] From the perspective of English's number system, these are all count nouns, not mass nouns.

[4] The cohesive function here is the ellipsis of the potential lexical repetition (ships).

did	you	see	the ships
Finite	Subject	Predicator	Complement

we	saw		the whole squadron of ships
Subject	Finite	Predicator	Complement

Misconstruing Location (Enhancement) As Possession (Extension)

Martin (1992: 305):

A further distinction can be drawn between alienable and inalienable part/whole relations.  These are not grammatically distinguished in English possessive constructions (e.g. the front garden's trees and the car's boot).  But when the whole is realised as a qualifier, inalienable possessives favour the preposition of, whereas alienable ones are more likely to be introduced with locative prepositions such as in, inside, within etc.

Table 5.6. Alienable and inalienable possession
alienable (the tree in the garden)	inalienable (the boot of the car)
garden:tree::	tree:boot::
room:chair::	room:wall::
stream:fish::	fish:gills::
car:spare tire::	car:boot::
cup:coffee …	cup:handle …

Blogger Comments:

[1] In SFL theory, composition (meronymy) and possession are two distinct subtypes of the expansion category extension (see Halliday & Matthiessen 1999: 146).  Here possession is presented as a subtype of composition.   That is, it misconstrues two SFL co-hyponyms as related by hyponymy.

[2] Here the relation of alienable possession (extension) is confused with the relation of location (enhancement).

Misconstruing Ellipsis Of Repetition As Relational Meronymy [2]

Martin (1992: 305):

Alongside this set [of items realising relational meronymy], there are two groups of items which function in what Halliday (1985: 174) calls Pre-Deictic and Pre-Numerative position.  Those in Pre-Deictic position name components with respect to their location in time or space (e.g. the top of the wall): top, inside, side, bottom, edge, middle, outside, environs, start, finish, beginning, end.  These items may function "cohesively" in text: We played well that set.  — Not at the start.

Those in Pre-Numerative position measure out some portion of the whole (e.g. a sip of beer): jar, bottle, schooner, glass, midi, jug, can, loaf, mouthful, spoonful, pound, ounce, kilo, yard, metre.  These are also used "cohesively": Would you like a beer? — Just a midi thanks. 

Blogger Comments:

[1] In SFL theory, this nominal group function is termed 'extended Numerative: facet', which is the cross-classification of 'partitive' and 'type'.  See Halliday & Matthiessen (2004: 333).

the	top	of	the	wall
Numerative: facet			Deictic	Thing

[2] Not all of these items necessarily represent parts of wholes.

[3] The cohesive function here is the ellipsis of the potential lexical repetition (set).

we	played		well	that set
Subject	Finite	Predicator	Adjunct	Complement

not	at the start of the set
mood Adjunct: polarity	circumstantial Adjunct

[4] In SFL theory, this nominal group function is termed 'extended Numerative: quantum', which is the cross-classification of 'quantitative' and 'measure'.  See Halliday & Matthiessen (2004: 333).

a	sip	of	beer
Numerative: quantum			Thing

[5] The cohesive function here is the ellipsis of the potential lexical repetition (beer).

would	you	like	a beer
Finite	Subject	Predicator	Complement

I	would	like	just	a middy of beer
Subject	Finite	Predicator	mood Adjunct: intensity	Complement

Misconstruing Ellipsis Of Repetition As Relational Meronymy [1]

Martin (1992: 305):

As with hyponymy, the relationship between part and whole may be named.  Thus a kind of chair is proportional with a part of the chair, but with the name of the relational [sic] filling a Pre-Numerative (Halliday 1985: 174) rather than a Pre-Classifier position.  The relational items may function cohesively between clauses, as in The chair's broken. — Which part?  The following items are among those realizing relational meronymy:

part, content, ingredient, fitting, member, constituent, stratum, rank, plane, element, factor, component, faction, excerpt, extract, selection, piece, segment, section, portion, measure

Blogger Comments:

[1] In SFL theory, this nominal group function is termed 'extended Numerative: portion', which is the cross-classification of 'partitive' and 'measure'.  See Halliday & Matthiessen (2004: 333).

a	part	of	the	chair
Numerative: portion			Deictic	Thing

[2] The cohesion in the response is ellipsis of the potential lexical repetition of chair (and of the Finite and Residue).

the chair	’s	broken
Subject	Finite	Complement

which part of the chair	’s	broken
Subject	Finite	Complement

The notion of a 'relational meronymy' therefore arises from not recognising a cohesive relation.

[3] Not all of these items necessarily represent parts of wholes.

Misconstruing Antonymy

Martin (1992: 301-2):

One further category that needs to be brought in here is antonymy.  Whereas synonyms are co-hyponyms for which differences in meaning may not matter, antonyms are co-hyponyms for which they must — because the items in question are opposed in meaning rather than complementing each other; 

Blogger Comments:

[1] Definitionally, synonyms are not types (hyponyms) of co-hyponyms.  In SFL theory, synonymy is a lexical means of creating cohesion in discourse.

[2] Definitionally, antonyms are not types (hyponyms) of co-hyponyms.  In SFL theory, antonymy is a lexical means of creating cohesion in discourse.

[3] Complementing is most generally concerned with completion, and often involves not just opposites (e.g. art: complementary colours), but contradiction (e.g. physics: particle-wave-field).  Halliday (2008: 84):

In the most general sense, complementarity is a special form of complexity; one can think of it perhaps as the management of contradiction.