Many English writing exercises ask learners to produce a full essay before they can reliably shape one sentence. Sentence combining reverses the scale. Give a learner three short statements, ask for one or two sentences that preserve every meaning, and the work moves to a manageable piece of language. This can sharpen syntactic choice, provided the lesson treats combining as revision rather than a race to make sentences longer. The intended reader is a Taiwanese senior-high or university learner whose writing is grammatically cautious: subject, verb, object. Full stop. Repeat. The problem shows up in GEPT and university assignments when every fact receives equal weight and the paragraph sounds like notes copied into complete sentences.
What the task asks
Consider these kernels:
The night market was crowded. We stayed for two hours. The rain began at nine.
Several answers are possible. ‘Although the night market was crowded, we stayed for two hours, until rain began at nine’ makes crowding a concession and rain the endpoint. ‘We stayed at the crowded night market for two hours. Rain began at nine’ keeps the facts more independent. A writer has to choose the relationship before choosing the connector. That choice is the lesson. Combining can practise subordination, relative clauses, apposition, coordination and punctuation, but the grammar should follow the intended emphasis. If a worksheet supplies although in advance, it has already made the most interesting decision for the learner.
Research on second-language sentence combining is less tidy than promotional summaries imply. An early ESL experiment gave systematic combining practice to first-year composition students. The treatment group improved on some measures, yet the prediction that they would write better compositions was not fully confirmed. The 1982 study remains useful because it separates success on the exercise from success in extended writing.
Think-aloud research used protocols with nine advanced second-language writers. Open combining tasks prompted more high-level planning than controlled tasks, and writers evaluated whether their constructions were appropriate. They paid little attention to cohesion across the larger text. The small sample rules out broad claims, though the study’s process evidence suggests that open tasks demand genuine decisions while isolated combining does not automatically teach paragraph flow.
The exercise and the paragraph remain different achievements.
Put combining after meaning
Useful lessons begin with a draft or a message, not ten unrelated sets of kernels. Ask learners to write a short response first. Then select two or three places where the sentence pattern obscures a relationship. The learner already knows what the passage is trying to say, so the new task is to recast that meaning. One example might be a company update written by a 27-year-old engineer in Hsinchu:
The shipment arrived on Tuesday. Two sensors were damaged. We can finish testing on Friday. We need replacements by Thursday afternoon.
Writers could group arrival with damage, then connect the replacement deadline to Friday’s test. A plausible revision is: ‘The shipment arrived on Tuesday with two damaged sensors. We can still finish testing on Friday if the replacements arrive by Thursday afternoon.’ The word if exposes the operational dependency. A longer sentence without that relationship would add weight without clarity. This is also why planning an English paragraph before writing the first sentence remains relevant. Paragraph planning decides which ideas belong together. Sentence combining decides how nearby ideas should relate on the page.
Use constraints that protect meaning
Good prompts give the learner room to choose and something firm to protect. Require every fact to survive. Set a maximum of two sentences. Ask for one version suitable for a message to a colleague and another for a formal report, then have partners explain which idea became background information, which remained in the main clause and how that choice changes what the reader notices first. Meaning can change quietly during combination. ‘Because Mei missed the train, she arrived late’ asserts a cause. ‘Mei missed the train and arrived late’ records sequence and invites the reader to infer. ‘Mei, who missed the train, arrived late’ places the missed train inside a relative clause. These are not decorative variants. Each tells the reader how to rank the information.
Keep a decombining check nearby. After writing the new sentence, split it back into its smallest claims. If one fact vanished or a new cause appeared, revise. I prefer this check to counting subordinate clauses. It catches the exact moment when syntactic ambition starts altering the message.
A short practice sequence
Use four rounds with uneven demands.
First, combine two related sentences without adding a connector. This invites a prepositional phrase, an adjective or apposition. Next, combine three statements in two different ways and underline the main clause in each. The third round starts from a learner’s own paragraph and permits deletion of one repeated idea. The final round asks the learner to reject combination where two short sentences create a deliberate beat. That last refusal belongs in the exercise. Short sentences can mark a turn, control reading pace or prevent a condition from disappearing inside a crowded structure. ‘The figures are provisional. Do not send them to the client’ should probably remain two sentences in an urgent workplace message.
One recent quasi-experimental study followed 40 EFL undergraduates for a semester and reported gains in mean T-unit length, clauses per T-unit and dependent clauses per clause for the combining group. Those measures capture syntactic complexity, and the study supports sustained classroom use. Its sample came from one programme, selection was purposive, and longer T-units still do not prove that ideas became clearer. The open report is evidence for development in sentence structure, with that boundary attached.
What to assess
Accuracy still decides.
Do not award the highest score to the answer with the most clauses. Check four things: every original meaning remains, logical relations are accurate, the sentence can be read without backtracking and the rhythm fits its purpose. Grammar accuracy matters, yet an accurate sentence that invents causality has failed the task. The teacher’s feedback can be compact. Circle the main clause. Draw a line from each dependent unit to the idea it modifies. Mark any relation that the source sentences did not justify. Then let the learner produce a second version with a different emphasis.
For tomorrow’s practice, take four consecutive sentences from an old draft. Write two combined versions, one that foregrounds a condition and one that foregrounds a result. Read both aloud. Keep the version whose emphasis matches what you meant, even if it is the shorter one.