Discourse grammar and cohesion

A sentence is a syntactic unit; a text is held together by cohesion — the grammatical and lexical ties that connect one sentence to the next. C2 reading and writing both depend on automatic processing of cohesion: tracking what a pronoun refers to three sentences back, recognizing when one substitutes for suspect, recognizing when did so substitutes for an entire verb phrase, noticing when a chain of synonyms (scholar / academic / researcher) builds lexical cohesion across a paragraph.

Russian and English share most cohesive devices but distribute them differently. Russian relies heavily on inflection (gender and case agreement do anaphoric work), null arguments (pro-drop in casual register), and demonstratives (тот, этот, такой). English, with poorer inflection and no pro-drop, relies more on explicit pronouns, substitution proforms (one, ones, do so, do it, so, not, this, that), and ellipsis. Russian-speaker C2 writing often shows residual under-use of these English-specific devices — too many full repetitions where a pronoun or substitution would carry the day.

This lesson covers (1) anaphoric (backward) and cataphoric (forward) reference, (2) substitution with one, ones, do so, do it, so, not, (3) ellipsis — what English permits to be omitted under recovery, (4) lexical cohesion through repetition, synonymy, hyponymy, and collocation.

Anaphora and cataphora — the directions of reference

Anaphora is reference backward to something already mentioned. Cataphora is reference forward to something about to be mentioned.

Anaphoric reference (backward)

• *The senator gave a speech. She then took questions. (She refers back to the senator)
• *Two protesters were arrested. They were released within hours.
• *Three texts shaped his thinking: the Republic, Leviathan, and On Liberty. These he reread every decade.

This is the default pattern. The referent is established first; the pronoun or demonstrative refers back to it.

Cataphoric reference (forward)

• When he arrived at the courthouse, the senator faced a wall of cameras.
• Before it could be enacted, the bill needed two more votes.
• This is what surprised me: she didn’t even mention the audit.

Cataphora is rarer but powerful — it creates suspense or rhetorical anticipation. The pronoun appears first; the referent resolves later.

When cataphora fails

• He cleared his throat. The senator began. (ambiguous — was he the senator or someone else?)

For cataphora to work, the resolution must be clear and reasonably close. If the gap between pronoun and referent is too large, the reader gets lost.

Substitution — proform replacement

English uses dedicated proforms to substitute for noun phrases, verb phrases, and clauses. These are different from pronouns: pronouns refer to entities; substitution proforms replace constituents.

Substitution with one and ones (NP-substitution)

• I want the blue one, not the red one. (one = car, shirt, whichever NP head)
• These books are mine; those ones are hers.
• He preferred a leather chair, but the cheaper ones were vinyl.

One substitutes for a previously mentioned count noun. Ones is the plural. They are not pronouns — they retain modifier slots (the blue one, the cheaper ones).

Substitution with do so, do it, do that (VP-substitution)

• She has not yet filed the brief, but she will do so tomorrow. (do so = file the brief)
• He promised to call, and he did. (bare did — the auxiliary substitutes)
• They asked her to resign; she refused to do so.
• He often skips breakfast. I never do.
• They wanted to investigate, and they have already done so.

Substitution with so and not (clause-substitution)

• *Will they accept? — I think so. (so = they will accept)
• *Will they accept? — I’m afraid not. (not = they won’t accept)
• *Did she really say that? — I believe so.
• *Are they likely to file an appeal? — I doubt it. (with it — slightly different pattern)

I think not in modern AmE sounds slightly formal or even theatrical; I don’t think so is the AmE default. I think so is universal.

Substitution with demonstratives this/that

• He resigned without notice. That surprised no one. (sentential — that = the entire previous proposition)
• She wrote three drafts. This I can confirm. (anaphoric, with object preposing)
• I have one piece of advice: this — read the contract before signing. (cataphoric)

Demonstratives substitute for entire propositions or events, often functioning as topic anchors at the start of new sentences.

Ellipsis — what English permits to be omitted

Ellipsis drops constituents that the reader can recover from the prior context. English permits ellipsis in three main places.

1. Verb phrase ellipsis

The VP is omitted after an auxiliary, leaving only the auxiliary as a placeholder:

• She will testify, but he will not [testify].
• I haven’t read the report, but my partner has [read the report].
• They asked me to resign, but I won’t [resign].
• He says he didn’t lie, and his lawyer believes he didn’t [lie].

The auxiliary stays; the verbal content drops. This is the most common English ellipsis pattern.

2. Gapping

In coordinated clauses, the verb can be dropped in the second clause:

• Maria called her brother, and Sarah [called] her sister.
• He ordered pasta; she [ordered] the fish.
• The Senate passed the bill in three days; the House [passed it] in three weeks.

Gapping is a written-prose device; it produces compression but requires clear parallelism.

3. Nominal ellipsis

The head noun is dropped, leaving modifiers:

• I prefer the red [shirt] to the blue [shirt].
• Among the proposals, the most ambitious [proposal] was hers.
• Of all my colleagues, the most ambitious is Maria.

Nominal ellipsis often co-occurs with one/ones substitution: the red one (substitution) or the red (ellipsis).

Ellipsis after as, than, like

• He works harder than I [work].
• She is as tall as he [is].
• He acted as the situation demanded [him to act].

In casual AmE, than and as take object-case pronouns: harder than me, as tall as him. Strict prescriptive style uses subject-case (than I, as tall as he). Both occur in C2 writing; the object form is now the AmE default.

Lexical cohesion — repetition, synonymy, hyponymy

Cohesion is not only grammatical. It is also lexical — the chain of related words that builds across a text. C2 writers manage these chains deliberately.

Direct repetition

The simplest cohesive tie: repeating the same word.

• The senator faced the committee. The senator was prepared for the committee. (repetition)

Direct repetition is heavy but powerful in emphasis. In short passages it can drum the topic into the reader’s awareness.

Synonymy

Replacing the repeated noun with a synonym:

• The senator faced the committee. The lawmaker had been preparing for weeks. The legislator opened with a quote from Lincoln.

Three lexically distinct labels for the same referent — common in journalism, where the rule “vary your word” is taught early. The risk: synonymy can confuse if the reader is uncertain whether the new label is the same referent.

Hyponymy

Replacing the repeated noun with a more general or more specific term:

• He carried a briefcase. The bag was leather. (briefcase = hyponym of bag)
• She trained as a journalist. The profession rewards persistence more than talent. (journalist = hyponym of profession)

This is the climb up the abstraction ladder: from specific to general, or vice versa. AmE essayistic prose climbs up and down this ladder constantly.

Collocation

Words that habitually co-occur in the same semantic field create cohesion even without direct reference:

• The patient was admitted overnight. The doctor ordered tests. The nurse adjusted the IV. (patient, doctor, nurse, IV — medical collocation network)
• The plaintiff filed the motion. The judge sustained the objection. The defendant declined to comment. (legal collocation)

C2 prose builds collocation networks unconsciously — and C2 reading parses them automatically. A breakdown in expected collocations is one of the strongest signals of non-native writing.

The cohesive chain across a paragraph

Read this passage with cohesive devices marked:

The senator [referent A] arrived late. She [pronoun → A] had spent the morning briefing her [genitive → A] staff. The lawmaker [synonymy → A] looked tired but composed. It [sentential → her arrival] surprised no one. They [pronoun → her staff] had already taken their [genitive → her staff] seats.

A single sentence pattern repeated through five different cohesive devices. This is the texture of well-managed AmE prose.

Reference vs ambiguity — when cohesion breaks

A cohesion chain breaks when reference becomes ambiguous:

• The senator argued with her chief of staff. She then resigned. (who resigned — the senator or the chief of staff?)

When pronoun reference is ambiguous, the writer must repeat the noun:

• The senator argued with her chief of staff. The chief of staff then resigned.

C2 writers manage anti-ambiguity actively — they pause on each pronoun and check that the antecedent is unambiguous.

It vs this vs that — the demonstrative-deictic split

Examples

• *The senator gave a speech. It lasted an hour. (it — neutral)
• *The senator gave a speech. This was the speech we had been waiting for. (this — bringing into close focus)
• *The senator gave a speech. That was the moment her career changed. (that — distancing, marking it as a past event)

The this/that split also handles forward and backward reference: Listen to this: he just resigned (cataphoric — close upcoming); He said he was leaving. That was a year ago. (anaphoric — distant past).

AmE notes

• Do so is more common in AmE writing than in AmE speech. He filed it, and his partner did so as well is journalistic; He filed it, and his partner did too is conversational.
• Sentential that is unmistakably AmE narrative rhythm: He arrived without warning. That changed everything.
• This as cataphoric topic-opener: Here’s the thing: this — they don’t trust the press. AmE conversation uses this this heavily.
• Bare auxiliary substitution is AmE conversational default: Did you call? — Yes, I did. / Will you come? — I will. / Have you read it? — I have.
• VP ellipsis is preserved in AmE writing: She will testify, but he won’t. The recovery is automatic for native readers.

Pragmatic note — when to repeat the noun

Sometimes a pronoun is technically unambiguous but the noun-repetition lifts register or emphasizes the entity:

• The chief justice opened the session. He read the opening summary. (neutral)
• The chief justice opened the session. The chief justice read the opening summary. (emphatic, formal — the title is being deliberately re-invoked)

Legal and ceremonial prose often re-uses titles where pronouns would suffice; this is a register marker, not a cohesion failure.

Common Russian-speaker mistakes

1. Over-repeating the noun where a pronoun suffices: The senator arrived late. The senator had spent the morning briefing the senator’s staff → *The senator arrived late. She had spent the morning briefing her staff. Russian permits repetition; English prefers pronouns.
2. Under-using do so/did so substitution: I asked him to leave, and he left → *I asked him to leave, and he did so. Do so is the AmE C2-writing register for VP substitution.
3. Wrong scope of so and not in substitution: Is the meeting cancelled? — Yes, that. → Yes, so I believe or Yes, I think so. That doesn’t substitute for clauses in this slot.
4. Calquing Russian zero-anaphora: Senator came. Said nothing. → *The senator came. She said nothing. English requires explicit subject in main clauses.
5. VP ellipsis without proper auxiliary: I will go, but he won’t (correct — auxiliary remains); I will go, but he (wrong — no auxiliary). The placeholder must remain.
6. Confusing this and that for proximal/distal reference: He resigned yesterday. This changed everything (close — the resignation is recent and in active focus); He resigned a decade ago. That changed everything (distant — the resignation is completed history). The proximal/distal split is real in AmE.
7. Reference ambiguity with multiple antecedents: She told her assistant that her keys were missing. Whose keys? Russian gender-agreement on her would disambiguate; English has no agreement. C2 writers recast: She told her assistant that the assistant’s keys were missing OR She told her assistant, “My keys are missing.”

Summary

• Anaphoric reference (backward) is the English default; cataphoric (forward) creates suspense and topic-anchoring.
• Substitution uses dedicated proforms: one/ones (NP), do so/do it/did (VP), so/not (clause), this/that (sentential).
• Ellipsis drops recoverable material: VP ellipsis (with auxiliary placeholder), gapping (verb in coordinated clause), nominal ellipsis (head noun).
• Lexical cohesion chains through direct repetition, synonymy, hyponymy, and collocation networks.
• This/that split: proximal vs distal in attention and time; controls topic continuity.
• C2 writers manage anti-ambiguity actively; cohesion that confuses is worse than cohesion that repeats.

Next lesson: Conjunctions and connectors — mastery — nevertheless, nonetheless, whereas, conversely, in light of, by virtue of, notwithstanding, albeit.