TOEIC Link Writing — Paraphrase Quality: Why Syntactic Transformation Beats Word-Swapping in the Integrated Task

The integrated reading-writing task asks candidates to restate ideas from a source in their own words, and the way a candidate does that restatement is one of the clearest quality signals a rater has. Most candidates paraphrase the easy way: they keep the source sentence's grammatical skeleton and swap the content words for synonyms. This produces a sentence that is technically not copied but is structurally identical to the source — and raters are trained to recognize it as low-quality paraphrase, close enough to borrowed language that it earns little credit. The high-band candidate does something different. They change the structure of the idea, not just its vocabulary, so that the restatement is a genuine re-expression rather than a lightly disguised copy. This is the difference between lexical and syntactic paraphrase, and it is worth a band.

This guide separates the two kinds of paraphrase, maps the five syntactic operations that produce real transformation, and lays out a training protocol. Because paraphrase quality is the engine of integrated-task performance, read this alongside evidence integration and source synthesis under the multi-source prompt, and because a paraphrase must still sit inside a well-built argument, situate it within claim-evidence-warrant paragraph construction.

The two kinds of paraphrase — and why raters can tell them apart

Lexical paraphrase changes the words while keeping the structure. Given the source The company reduced its operating costs by closing three regional offices, a lexical paraphrase produces The firm lowered its operating expenses by shutting several regional branches. Every content word has been swapped, and yet the sentence is structurally a clone — same subject-verb-object frame, same adverbial-of-means clause in the same position. A rater reading it recognizes that the writer has processed the vocabulary but not the idea.

Syntactic paraphrase changes the structure so the idea is repackaged: Closing three regional offices allowed the company to cut its operating costs. Here the means clause has become the subject, the causal relationship is now carried by the verb allowed, and the reader can see that the writer understood the relationship between the closures and the savings well enough to re-frame it. That comprehension is exactly what the integrated task is designed to measure, and syntactic transformation is the evidence of it. Synonym-swapping demonstrates a thesaurus; structural change demonstrates understanding.

The five syntactic transformation operations

Real paraphrase draws on a small toolkit of grammatical moves. Five operations cover the large majority of cases.

Nominalization and de-nominalization. Convert a verb into a noun or the reverse to shift the sentence's center of gravity. Prices increased becomes the increase in prices; the reduction of staff becomes staff were reduced. Changing what is packaged as the noun and what as the verb restructures the whole clause.
Voice change. Move between active and passive to relocate the emphasis and change the subject. Regulators approved the merger becomes the merger was approved by regulators. Voice change is the fastest structural transformation, but it must be motivated by emphasis rather than used mechanically.
Clause-type conversion. Turn a subordinate clause into a main clause or the reverse, or convert a relative clause into a participial phrase. Because demand fell, the factory closed becomes falling demand forced the factory to close — a causal subordinate clause collapsed into a subject-plus-verb.
Word-class shift with re-framing. Change the part of speech of the key idea and rebuild the clause around it. The policy was successful becomes the success of the policy becomes the policy succeeded in — each shift forces a different sentence frame.
Relationship re-expression. Restate the logical relationship using a different connective structure. A concessive although X, Y becomes Y despite X becomes X, yet Y. The information is identical; the grammatical packaging is not.

Combining two operations on a single source sentence produces paraphrase that no rater will mistake for word-swapping.

The trap: over-transformation that distorts meaning

Syntactic transformation is powerful enough to be dangerous. A candidate who restructures aggressively can change the meaning without noticing — reversing a causal direction, dropping a hedge that the source included, or making a qualified claim absolute. The integrated task penalizes meaning distortion more heavily than weak paraphrase, because a distorted restatement is worse than a faithful clone.

The discipline is transformation-then-verification: after restructuring, read the paraphrase against the source and confirm that the propositional content is identical — same claim, same strength, same causal direction, same qualifications. Structural freedom is licensed only by meaning fidelity. A paraphrase that reads as genuinely re-expressed but says something the source did not is a net loss.

The four-week paraphrase protocol

Paraphrase quality is built by drilling the operations until they are automatic under time pressure.

Week one — operation isolation. Take single source sentences and produce five paraphrases of each, one using each operation. The point is not to choose the best but to make each transformation available on demand, so that under pressure you are selecting from a toolkit rather than reaching for synonyms.

Week two — meaning-fidelity checking. Paraphrase source sentences, then audit each against the original for the four fidelity dimensions — claim, strength, causal direction, qualification. Build the reflex that pairs every transformation with a verification pass.

Week three — combination under time. Paraphrase to a clock, requiring at least two combined operations per source sentence. Speed forces the operations to become instinctive rather than deliberate.

Week four — full integrated responses. Write complete integrated-task responses, then have every source-derived sentence checked for both transformation quality and meaning fidelity. The endpoint is paraphrase that is structurally genuine and propositionally exact, produced at task speed.

The paraphrase-quality checklist

For every sentence you draw from a source:

Structure changed, not just words — at least one syntactic operation applied, not synonym-swapping.
Two operations for key sentences — combine nominalization, voice, clause-type, word-class, or relationship re-expression.
Meaning verified — claim, strength, causal direction, and qualifications identical to the source.
No over-transformation — the restructuring has not distorted or absolutized the original.
Integrated, not stitched — the paraphrase reads as your sentence, not a rebuilt version of theirs.

Why this separates bands

Two candidates can both avoid copying and still land a band apart, because one swaps synonyms into the source's frame and the other rebuilds the frame. The rater reads the first as vocabulary work laid over borrowed structure and the second as genuine comprehension expressed in the writer's own grammar. The integrated task exists to measure whether a candidate can take an idea from a source and make it their own, and syntactic transformation is the only paraphrase that proves they can. Word-swapping shows you own a thesaurus; restructuring shows you understood the idea — and only the second earns the top of the band.