Adapted from Discourse Analysis for Language Teachers by Michael McCarthy

(This book is available in the library. This is a summary for chapters 2 and 3).

Discourse and text interpretation

The grammar of English offers a limited set of options for creating surface links between the clauses and
sentences of a text known as cohesion. Basically, most texts display links from sentence to sentence in
terms of grammatical features such as pronominalization, ellipsis (the omission of otherwise expected
elements because they are retrievable from the previous text or context) and conjunctions of various kinds.
The cohesive items are clues or signals as to how the text should be read. Cohesion is only a guide to
coherence, and coherence is something created by the reader in the act of reading the text. Coherence is the
feeling that a text hangs together, that it makes sense, and is not just a jumble of sentences. The sentences
“Clare loves potatoes. She was born in Ireland.” are cohesive (Clare/She), but are only coherent if one
already shares the stereotype ethnic association between being Irish and loving potatoes, or is prepared to
assume a cause-effect relationship between the two sentences. So cohesion is only part of coherence in
reading and writing, and indeed in spoken language too, for the same processes operate there. Cohesive
markers create links across sentence boundaries and pair and chain together items that are related (e.g. by
referring to the same entity). But reading a text is far more complex than that: we have to interpret the ties
and make sense of them. Making sense of a text is an act of interpretation that depends as much on what we
as readers bring to a text as what the author puts into it.

Grammatical cohesion and textuality

1. Reference (This includes the pronoun it, discussed at the end of these notes.)
Reference items in English include pronouns (e.g. he, she, it, him, they, etc.), demonstratives (this, that,
these, those), the article the, and items like such a. Referents can be confirmed by looking back in the text;
this is called anaphoric reference. References to assumed, shared worlds outside of the text are exophoric
references. Because they are not text-internal, they are not truly cohesive. Forward-looking or cataphoric
reference often involves pronouns but it can involve other reference items too, such as the definite article. Its
most characteristic function is to engage and hold the reader’s attention with a “read on and find out”
message. Cataphoric reference is the reverse of anaphoric reference and is relatively straightforward.

2. Ellipsis and substitution SWAN 177-182

Ellipsis is the omission of elements normally required by the grammar which the speaker/writer assumes are
obvious from the context and therefore need not be raised. This is not to say that every utterance which is
not fully explicit is elliptical; most messages require some input from the context to make sense of them.
Ellipsis is distinguished by the structure having some “missing” element. If two people have to stack and
label a pile of items and one says to the reader “you label and I’ll stack,” the fact that label and stack are
usually transitive verbs requiring an object in the surface structure is suspended because the context
“supplies” the object. We shall concentrate here on the type of ellipsis where the “missing” element is
retrievable verbatim from the surrounding text, rather in the way that anaphoric and cataphoric references
are, as opposed to exophoric references. For example:
The children will carry the small boxes; the adults, the large ones.

Where “will carry” is supplied from the first clause to the second. English has broadly three types of ellipsis:
nominal, verbal and clausal. Nominal ellipsis often involves omission of a noun headword:
Nelly liked the green tiles; I myself preferred the blue.

The Romance and Germanic languages have this kind of nominal ellipsis and it should not present great
difficulties to speakers of those languages learning English.

Ellipsis within the verbal group may cause greater problems. Two very common types of verbal-group
ellipsis are also called echoing. Echoing repeats an element from the verbal group:
A: Will anyone be waiting? B: No, but she will one day, I’m sure.
B: Jim will, I should think. Varying degrees of ellipsis are possible:
Contrasting is when the auxiliary changes: A: Should anyone have been told?
A: Has she remarried? B: John should.
1
 should have. should have been.

With clausal ellipsis in English, individual clause elements may be omitted; especially common are subject-
pronoun omissions (“doesn’t matter”, “hope so”, “sorry, can’t help you”, etc.). Whole stretches of clausal
components may also be omitted:
He said he would take early retirement as soon as he could and he has.

Substitution is similar to ellipsis, in that, in English, it operates either at nominal, verbal or clausal level.
The items commonly used for substitution in English are:
One(s): I offered him a seat. He said he didn’t want one.
Do: Did Mary take that letter? She might have done.
So/not: Do you need a lift? If so, wait for me; if not, I’ll see you there.
Same: She chose the roast duck; I chose the same.

Where the speaker does wish to give prominence to the substitute do, then so is used as well:
I went to lock the gate. When I got there, I found somebody had already DONE so.

3. Conjunction SWAN 510

We include conjunction here in our discussion of grammatical contributions to textuality even though it is
somewhat different from reference, ellipsis and substitution. A conjunction does not set off a search
backward or forward for its referent, but it does presuppose a textual sequence, and signals a relationship
between segments of the discourse. In fact it is not at all easy to list definitively all the items that perform
the conjunctive role in English. Single-word conjunctions merge into phrasal and clausal ones, and there is
often little difference between the linking of two clauses by a single-word conjunction, a phrasal one, or a
lexical item somewhere else in the clause (e.g. consequently, as a consequence, as a consequence of).

Lexical cohesion
Related vocabulary items occur across clause and sentence boundaries in written texts across act, move and
turn boundaries in speech and are a major characteristic of coherent discourse. It is debatable whether
collocation properly belongs to the notion of lexical cohesion, since collocation only refers to the probability
that lexical items will co-occur, and is not a semantic relation between words. Here, therefore, we shall
consider the term “lexical cohesion” to mean only exact repetition of words in creating textuality, that
property of text which distinguishes it from a random sequence of unconnected sentences. We shall
consequently ignore collocational associations across sentence boundaries as lying outside of these semantic
relations.

Reiteration means either restating an item in a later part of the discourse by direct repetition or else
reasserting its meaning by exploiting lexical relations. Lexical relations are the stable semantic relationships
that exist between words and which are the basis of descriptions given in dictionaries and thesauri: for
example, rose and flower are related by hyponymy; rose is a hyponym of flower. Eggplant and aubergine
are related by synonymy (regardless of the geographical dimension of usage that distinguishes them). An
awareness of the usefulness of learning synonyms and hyponyms for text-creating purposes is central.

Vocabulary and the organising of text

A distinction is often made between grammar words and lexical words in language. This distinction also
appears sometimes as function words versus content words, or empty words versus full words. The
distinction is a useful one: it enables us to separate off those words which belong to closed systems in the
language and which carry grammatical meaning, from those that belong to open systems and which belong
to the major word classes of noun, verb, adjective and verb. This, that, these and those in English belong to a
closed system (as do the pronouns and the prepositions) and carry the grammatical meaning of
“demonstratives”. Monkey, sculpture, noise and toenail belong to open-ended sets, which are often thought
of as the “creative” end of language. In between these two extremes is another type of vocabulary that has
recently been studied by discourse analysts, a type that seems to share qualities of both the open and the
closed-set words. These words do a bit of lexical work (they are not as “empty” as grammar words are often
said to be), but, in another sense, we need to seek elsewhere in the text for their content, what we shall call
their lexicalisation. We shall call words such as issue, problem and assessment discourse-organising words,
2
since it is their job to organise and structure the argument, rather than answer for its content or field. They
are examples of the phenomenon called signalling. As well as representing text-segments, some of the
discourse-organising words additionally give us indications of the larger text-patterns the author has chosen,
and build up expectations concerning the shape of the whole text. If the discourse-organising words are seen
as signals of the author’s intent, then inability to understand them or misinterptretation of them could cause
problems.

But just how many words are there in a language like English? Some linguists have attempted to provide
answers, but probably no one has compiled a complete list. Here is a selection:

analysis approach assessment attitude belief classification

comparison concept consideration deduction diagnosis distinction
dogma evaluation evidence examination fabrication fantasy
hypothesis idea illusion inference insight interpretation
investigation misinterpretation notion opinion perspective
picture position reasoning recognition reflection scenario
speculation supposition tenet theory view viewpoint
vision

These words are also called anaphoric/cataphoric nouns. It might be worth reminding ourselves that
discourse-organising words operate predictively in text as well as retrospectively: if a discourse organiser
does not already have its lexicalisation in the earlier text we expect it to come later in the text and are on the
lookout for it, at least the efficient reader is.

Decide whether these sentences are examples of coherence or a certain cohesive device:
This shirt belongs to you. Always has. Always will.
His scholarship is unquestioned. His ideas are.
Come here at four or else.
Students who take notes by hand outperform students who type, and more type these days, studies show.
What am I looking at? My dinner or the … (I will dictate the missing word/s)?
Buy it. Read it. Enjoy it. // Enjoy it. Read it. Buy it.
Elon Musk fears that the development of artificial intelligence, or AI, may be the biggest existential threat
humanity faces.
A: The phone is ringing. B: I’m in the bath.
Quentin knew Julia and James better than he knew anyone else in the world, not excluding his parents, and
they knew him.
Some of the characters watch a TV show called “Space Junk,” which is practically an emblem of the novel;
it’s about some aliens who kidnap random earthlings on the assumption that they must be related and then
try to figure out why they behave the way they do.
The approval of a new democratic constitution took place by a national referendum on June 28, 1992, and it
entered into force on July 3, 1992.
Ms Sandberg writes that the most important career choice a woman makes is whom she marries.
The doctor whose wife has been reported missing is in his house. Police are digging in the garden.
I agonized over which tie to wear. I couldn’t decide between the dark conservative tie or a bright one.
Eventually I narrowed the choice down to a plain navy knitted job from Marks & Spencer and an Italian silk
number hand-painted in orange and red.
The aim is to cut the expense of renting and staffing branches from something like 60% of the total cost of
running a retail bank to 40%.

3
The pronoun it Swan 446-447

1 Anaphoric it: it refers back to something previously mentioned.

Where’s the horse? It (the horse) is in the stable.
It can stand for a whole sentence: You saved my life; I shall never forget it.

2 Anticipatory it: also called cataphoric it. It can anticipate:

a. an infinitive: It is wrong to lie.
b. an –ing form: It’s nice seeing you again.
c. a noun clause: It is a pity that you can’t come.

3 Unspecified it: in a great many cases it is used in such a way that it is not possible to point
to something specifically referred to.
a. Weather it: It is cold today.
b. Time it: It is four o’clock.
c. Space it: It is a long way to San Francisco
d. Idiomatic it: If you’re found out, you’ll catch it. If you go camping, you’ll have to rough it. Take
it as read. She’s got it. Have it in for her. He let him have it. You won’t hear the end of it. He’s
had a thin time of it since he lost his job. I don’t have it in me. You’ve got it made. You’ll never
hear the end of it. She let him have it (attack physically or verbally). They want to stick it to
the wealthy. We’ve never had it better. You either have it or you don’t. you’ll never hear the end
of it.
e. Unknown person It: Someone entered the room. It was a burglar.

4 Emphatic it: It was his mother who said that (not anybody else!). (The resulting structure is
also known as ‘cleft sentence’). It was in the kitchen that they were talking (that’s why you didn’t see
them). It was on my birthday that you called me a liar! (I remember well!) Swan 130-131

5 Objective it: I’d appreciate/prefer it if you left. I find it interesting that you like my music. Rumour
has it that you have a new job. He made it a problem to walk. I thought it strange that she would be angry. I
take it that you are leaving. We owe it to society to make this a better place. I’d love it if you could help me.

Swan has other entries that mention similar uses of it: 291.4, 428−429, 590.