3.1 Sentence Builds
Key Takeaways
- Sentence Builds (Part D) plays three scrambled word groups that you reassemble into one grammatical sentence and say aloud once
- The task feeds Sentence Mastery most heavily, plus Fluency and Pronunciation
- Use a fixed method: find the subject, attach the verb chunk, then place time/place/reason phrases at the end
- Restarting mid-sentence or rehearsing chunks aloud lowers your Fluency score and wastes the response window
- You must use the chunks exactly as given — do not add, drop, or change words
What the Sentence Builds Task Asks You to Do
Sentence Builds is Part D of the Versant English Test. You hear three short word groups (call them "chunks") played in a scrambled, random order, and your job is to rearrange them into one grammatically correct, sensible sentence and say that whole sentence aloud one time. You are not given the words on screen — you hear them, hold them in memory, and reorder them by ear.
A worked example. Suppose the machine plays these three chunks:
"was late again" ... "because of traffic" ... "the delivery driver"
The only ordering that produces a natural English sentence is:
"The delivery driver was late again because of traffic."
Notice what the task is really testing: you had to identify the subject (the delivery driver), attach the verb phrase (was late again), and place the reason clause (because of traffic) where English grammar wants it — at the end. That is exactly why this task feeds your Sentence Mastery subscore most heavily, and also contributes to Fluency and Pronunciation, because you have to say the finished sentence smoothly and clearly.
A Reliable Three-Step Method
Because you only get a few seconds, guessing at random permutations wastes time. Use a fixed procedure:
- Find the subject. Which chunk names a person or thing that can do or be something? That chunk almost always starts the sentence.
- Attach the verb chunk. Which chunk contains the main verb? Put it right after the subject to form the core Subject-Verb-Object spine.
- Place the extra chunk. Time phrases (by Friday), place phrases (to the airport), and reason phrases (because of traffic) are modifiers — they usually go at the end.
Two more original examples show the pattern:
- Chunks: "a taxi" / "we took" / "to the airport" → "We took a taxi to the airport." (Subject we, verb took, object a taxi, place phrase last.)
- Chunks: "by Friday" / "the report" / "needs to be finished" → "The report needs to be finished by Friday." (Subject the report, verb phrase needs to be finished, time phrase last.)
| Chunk role | Question it answers | Typical position |
|---|---|---|
| Subject noun phrase | Who or what? | Start |
| Verb / verb phrase | Does or is what? | After subject |
| Object / complement | Whom or what receives it? | After verb |
| Time / place / reason phrase | When, where, why? | End |
Handling Longer, Harder Builds
Not every set is a tidy three-word sentence. Sometimes a chunk contains a subordinate clause or a connector, and the difficulty jumps. Consider:
"if the weather is bad" / "will be postponed" / "the outdoor event"
Here one chunk is a full if-clause. The subject is the outdoor event, the main verb phrase is will be postponed, and the condition attaches at the front or the end. Both "The outdoor event will be postponed if the weather is bad" and "If the weather is bad, the outdoor event will be postponed" are correct — so you can choose whichever you can say most smoothly. When a chunk starts with a connector such as if, because, when, although, or that, treat that chunk as a movable modifier and let the subject-plus-verb core anchor the sentence.
A useful ear-training habit is to listen for the chunk that cannot stand alone. A fragment like "because of the delay" or "that we discussed" is clearly a modifier or a relative clause, so it is never the subject — it clips onto the core. Identifying the dependent chunk first often makes the whole ordering fall into place.
One more point specific to the machine: the ASR is scoring your spoken output, so a build that is grammatically valid but that you deliver haltingly still costs Fluency. It is better to produce a slightly simpler valid ordering fluently than to attempt the most elegant ordering and stumble. Accuracy and smooth delivery are scored together, so aim for the sentence you can say cleanly on the first pass.
Traps That Cost Points
- Restarting mid-sentence. If you begin, stop, and begin again, the ASR hears hesitation and your Fluency drops. Decide the order silently first, then say the sentence once, cleanly.
- Rehearsing chunks out loud. Whispering "was late... the driver... no wait..." burns your response window, and the microphone may capture the wrong, half-formed version. Think silently; speak only the final answer.
- Ignoring meaning. Sometimes two orderings are grammatically possible but only one makes sense. Chunks "the manager" / "the email" / "sent" could form "The manager sent the email" or "The email sent the manager." Only the first is sensible — real-world logic breaks the tie.
- Adding or dropping words. You must use the chunks as given. Do not insert extra words or leave one out to make the sentence "easier."
How to Drill It
Practice with a partner or a text-to-speech app: have three chunks read to you in random order and rebuild the sentence within four seconds. Start with short six-to-eight-word sentences and build up to longer ones with a subordinate clause. Because the same syntax patterns (SVO plus a trailing modifier) repeat constantly, ten reps a day quickly makes the reordering automatic — which is the whole point, since on test day you want the grammar to feel like reflex, not calculation.
You hear these three word groups in random order: "before the meeting" / "the agenda" / "please review." What is the correct rebuilt sentence?
According to the recommended three-step method, which chunk should you identify first when rebuilding a Sentence Build?
Why is restarting a sentence partway through especially costly on the Sentence Builds task?