Topic 4: Writing cohesively

COM111 Professional writing

Reference | Ellipsis | Conjunctive relations | Lexical relations | Summing up 

Lexical relations 

These are relations between the lexical items, or meaning-bearing words, of a text.  In looking for lexical relations, we look at the word choice in texts and observe the connections of meaning between words.  These connections of meaning (semantic connections) are a very important way in which a text sticks together.  Semantic connections are different from grammatical connections.  In your work on Chapters 4 and 5 of Writing in Australia, you were focusing on grammatical connections.  For example, in 'The boy who had the ball threw it', relations between 'boy', 'ball' and 'threw’' are all grammatical relations: between subject and verb; between subject, verb and object.  But there are also links between words independent of the relations set up by the grammar of the English language.

Halliday helpfully refers to lexical structure as a 'map' of a text.

The systemicists' way of analysing lexical relations is through Lexical strings

These are patterns of connection between the meaning rather than the grammatical status of words.  In working out the lexical strings in a given text, systemicists look for such things as:

  • Repetition 
  • Synonyms and antonyms 
  • Words we expect to go together 

Producing a full-blown lexical string as the systemicists would is a bit complicated and scary, so we'll do something simpler.  But first, here is what a formal lexical string looks like.  It is Eggins' lexical string for the text from which came the remarks, 'But in Switzerland they give you cognac.  Here they give you tea and bickies' we looked at in Topic 3.  You haven't seen the full text but I want you to see what a lexical string looks like, remembering that you won't be asked to do anything as complex as this:



Source: Eggins p. 104, Figure 4.5



Eggins, p. 103, defines the lexical string as follows:

A lexical string is a diagram of all the lexical items that occur sequentially in a text that can be related to an immediately prior word either taxonomically or through an expectancy relation.


And here is Eggins' explanation of how you produce it:

Words which are related taxonomically to each other are lined up under each other, to give a vertical string.  The words are linked by a straight line, on which we write the type of relationship between each two words on the string (e.g. synonymy, repetition).  Words which relate through expectancy are written slightly to the right of the vertical line, and are joined with a labelled diagonal line.  We do not display every lexical item in a text but only those which form a chain (minimally two words at different points in the text).  Usually , however, we concentrate on displaying only the main lexical strings (those with at least three or four words in them). (p. 103).


As such, the lexical string is an important tool for revealing the cohesion of a text.  Even without constructing a formal lexical string, though, we can make good use of the principle of lexical organisation to discern how texts stick together and make sense in quite creative ways.

Analysing lexical organisation can really help expose the patterns of cohesiveness in a text—such as a poem—which does not always connect words tightly through grammatical relations.  Here's a little piece of analysis produced by a poet and literary critic which shows us how we can talk about lexical connections helpfully without actually producing a lexical string.  Hazel Smith is analysing a poem by Sylvia Plath.  Smith uses the traditional literary terms of metaphor, synecdoche and so on, but in a footnote comments that what she is analysing is what linguists would call lexical strings.  Synecdoche is a figure of speech in which relationships between parts and wholes, and relationships between members of the same class, are set up.  As we shall see presently, these kind of classificatory relationships are an important aspect of lexical organisation.

Hyperscapes in the Poetry of Frank O'Hara
When a poem comprises several different metaphors these usually nod in the direction of an overall signified. But, at the same time, the synecdoches which form the basis of the metaphors also forge their own lines or chains of association which thread through the poem.3 We can see this process at work in Sylvia Plath's poem 'Morning  Song' (Plath 1981, pp. 156-57):
Love set you going like a fat gold watch.
The midwife slapped tour footsoles, and your bald cry
Took its place among he elements.

Our voices echo, magnifying your arrival. New statue.
In a drafty museum, your nakedness
Shadows our safety. We stand round blankly as walls.

I'm no more your mother
Than the cloud that distills a mirror to reflect its own slow
Effacement at the wind's hand.

All night your moth-breath
Flickers among the flat roses. I wake to listen:
A far sea moves in my ear.

One cry, and I stumble from bed, cow-heavy and floral
In my Victorian nightgown.
Your mouth opens clean as a cat's. the window square

Whitens and swallows its dull stars. And now you try
Your handful of notes;
The clear vowels rise like balloons.
In this poem a plethora of metaphors (such as the child as a new statue in a museum, or the mother as 'cow-heavy and floral') contribute to the overall signified of the baby's arrival and the emotions of the parents. However, intratextual lines of association also run between the metaphors, constituted of part/whole relationships or members of the same class. For example, we can trace lines linked by sound (echo, notes, cry); emptiness (blankly, effacement) and parts of the body (mouth, footsoles, hand). It is also possible to pull some of the other more distant elements in the poem into these chains, for example, the gold watch in 'Love set you going like a fat gold watch' could be worn on the hand. these threads then intersect into a web of association within the poem which pulls against the unifying and externalising claims of metaphor. This contributes to the multilayered effect of the poem, and reinforces on a structural level the tension between the centred emotions of joy, and the decentred feelings of disturbance, experienced by the patents.

3. In discourse analysis these would be referred to as lexical strings. See Eggins, 1994, p. 103.

Source: Smith, H. 2000, Hyperscapes in the poetry of Frank O'Hara: Difference/homosexuality/topography, Liverpool University Press, Liverpool, pp. 83-84.




Hazel Smith's analysis of 'Morning Song' suggests how lexical organisation is a key aspect of successful creative writing: the trick is to invigorate your reader by setting up fresh, not entirely predictable, relationships between words, while taking into consideration your reader's need not to be totally bamboozled by the elements you are putting together.  So, as writers we exploit the tendency in Western thinking to organise things into classes or categories, which means that we tend to think of certain things as sub-classes of others.  If I said to you, 'the doctor thought my brother had glandular fever but it turned out to be chronic fatigue syndrome', and expected you—as I would—immediately to understand what I was saying, I would be exploiting the fact that you realise that glandular fever and chronic fatigue syndrome are both sub-classes of the class, 'illness'.  You would know that I was not discussing my brother's difficulties in meeting his mortgage repayments or settling with a neighbour whose dog had eaten his prize-winning collection of dahlias.  If, on the other hand, you read the following:
There was a fever on his brow
A dog was eating dahlias
The bailiffs were at the door

you might feel that this was a desperate attempt to start a postmodern, avant-garde poem, and that you had absolutely no desire to read on.  It takes a sure hand to scramble classificatory associations between words without making your writing descend into gobbledy-gook.

In addition to classificatory lexical relations, both writers and readers depend on relations between words we can expect or predict, even if they are not as tight as the classificatory relation between oak and tree.  Suzanne Eggins calls these kind of patterns 'relations of expectancy'.  If I said to you, 'my brother has taken the neighbour's dog to the sportswear store to have it put down because it has eaten all his dahlias' you'd think there was something intellectually as well as morally wrong with my brother.  But if I substituted 'vet' for 'sportswear store' you'd probably still think my brother wasn't the kind of person you'd want as a neighbour but not doubt his grasp of the way the world worked.  (I feel I should add at this point that my brother enjoys excellent health, has paid off his house, grows Australian native plants, not dahlias, and adores dogs).  The relation between 'dog' and 'have put down', on the one hand, and 'vet' or 'sportswear store' on the other is one of expectancy.  When we hear about a dog being out down we expect the place where it happened to be a vet's.

 Drawing together what I have just covered discursively, Suzanne Eggins says that we can observe two main kinds of lexical relations between words, taxonomic relations and relations of expectancy:

A. Taxonomic relations
in taxonomic relations (relations of classification) words are related to each other on a class/sub-class basis:
For example, in the first paragraph of Dirt Music, Pain Clark is a sub-class of boat

We include in these kind of relations:

  1. Contrast
    • hot-mild
    • a fast pace-a leisurely pace

  2. ii) Similarity
    Here we note both synonyms and antonyms (antonyms are opposites but logically rely on the principle of similarity)
    • sickness-illness, sad-unhappy
    • sickness-health, sad-happy

    and repetition of the same word or phrase (in other words the items are totally similar) 
I do not like thee, Dr Fell
The reason why, I cannot tell
But this I know, and know  full well
I do not like thee, Dr Fell

In relations of taxonomy we also note whole-part relations and relations of parts to a common whole.  Refer back to Hazel Smith's discussion of these relations in her analysis of Plath's 'Morning Song'.

Some other examples:
server web-site to computer (from Dirt Music, paragraph 1)
apples to pears (both belong to class of 'fruit')
chairs to desks (both belong to class of 'office furniture')
sail to boat ('sail' a part of 'boat')

And here is Richard Hinds exploiting taxonomic relations in his suggestion for a quiz sports players should be made to do, post the Shane Warne affair:
4. Some forms  of diuretics are illegal because they can mask the presence of:
a) baked beans
b) A second chin
c) Andy Bichel
d) Anabolic steroids

11. Drug tests may also show the use of certain recreational drugs, although not all of these are banned under the ACB drug code:
a) Mexicana
b) Capricciosa
c) Super Supreme
d) Marijuana


Source: Sydney Morning Herald, March 1-2, 2003, p. 74.


B. Relations of expectancy

This is a relation between a word and another we expect might go with it.

For example, in the first paragraph of Dirt Music, after reading that Georgie had traipsed through the Uffizi, we are not surprised (we could have predicted) the reference to a 'footsore tourist'.  The process 'traipsed' refers to movements with the feet so we can expect—and not be thrown by—a reference to 'foot' a little later on in the text.  There is thus a relation of expectancy between traipsed and footsore tourist.

Expectancy relations can be between an action and a typical 'doer' of the action, like 'traipsed' and 'footsore tourist'.  Other examples would be:
operate-surgeon
paint-artist
They can also be between a process and a participant affected by the process*:
saddle-horse
drive-car
They can also be between an event or process and the location where it takes place.  Refer back to my example of where my brother took the neighbour's dog.
Some other examples:
teach-school
eat-restaurant
They can also be between words which, as Halliday, p. 310, says, 'have a more than ordinary tendency to co-occur'.  We call this collocation.
Some examples:
pepper and salt
girls were made to love and kiss (as the song says—don't blame me!)
war and peace
So these relations of taxonomy (classification) and expectancy are what systemicists map when they construct Lexical strings

More simply, though, we can extract the principle of lexical organisation and just list instances of things like repetition, synonyms and antonyms, words which we can see are in a relation of expectancy with prior words in a stretch of text, and so on.  For the last Activity for this week's work, you will see that I have asked you to have a go at giving an account of the lexical organisation of the descriptive passage you yourself have written.  Look back at Hazel Smith's analysis of the Plath poem, 'Morning Song', at what I have said in this section, and at the summary of taxonomic and expectancy relations, and see what you can manage to pull out from your own text.  You don't have to be exhaustive—we're just keen to see that you are able to grasp the principle of lexical organisation and to discern some of the relevant word patterns.

*See Topic 7 for more on processes and participants

Reference | Ellipsis | Conjunctive relations | Lexical relations | Summing up