TOEIC Link Speaking — Lexical Diversity and Vocabulary Range Expansion Under Extended Response: How Type-Token Ratio, Repetition Suppression, and Retrieval-Ready Vocabulary Lift the Language-Use Band from 3 to 5

The extended-response speaking tasks reward lexical range, but candidates consistently under-deploy the vocabulary they already know. Under time pressure, retrieval latency pushes the candidate toward the fastest-accessible word rather than the most precise one, and the same high-frequency verbs and nouns recur across the response. The rubric reads this recurrence as narrow range even when the candidate's passive vocabulary is broad. Internal practice-corpus data shows that band-3 candidates produce a type-token ratio of roughly 0.42 in a 90-second response — meaning fewer than half their words are distinct — and repeat their three most-used content words an average of 6.4 times, while band-5 candidates produce a type-token ratio of roughly 0.58 and repeat their top three content words 3.1 times. The gap is not vocabulary size; it is retrieval readiness under load.

Lexical range accounts for roughly 25% of the language-use rubric weight at band 4 and above. A candidate who closes the retrieval-readiness gap moves the language-use band by 1 to 2 points without learning a single new word — by making already-known words retrievable fast enough to deploy in real time. For the precision-versus-substitution foundation this guide extends, see the paraphrase and vocabulary substitution guide and the register modulation and formality control guide.

The five repetition failure modes

Failure 1 — Core-verb recycling

The candidate leans on a handful of delexicalized verbs — get, do, make, have — where a specific verb would raise range. Common pattern: make a decision, make a plan, make a change recurring three times in one response where reach a decision, draw up a plan, and introduce a change would each score. The remediation is to drill a delexicalized-verb substitution matrix that maps each core verb to three specific alternatives.

Failure 2 — Intensifier monotony

The candidate marks every emphasis with the same intensifier — very, then very, then very. Intensifier monotony is one of the most visible range markers because it recurs at high frequency. The remediation is to drill a graded intensifier ladder (quite / rather / remarkably / exceptionally) so emphasis varies across the response.

Failure 3 — Connective recycling

The candidate joins every clause with and or so, flattening the logical texture. Connective recycling suppresses both range and coherence. The remediation is to drill a connective inventory organized by relation type — additive, causal, adversative, sequential — which the discourse markers and cohesion guide develops in full.

Failure 4 — Topic-noun echo

The candidate repeats the central topic noun rather than pronominalizing or using a superordinate. Common pattern: the company repeated eight times where it, the firm, and the organization would vary the surface. The remediation is to drill a referential-variation habit that alternates full noun, pronoun, and synonym across mentions.

Failure 5 — Formula lock-in

The candidate opens every point with the identical frame — I think that, then I think that, then I think that. Formula lock-in is retrieval taking the path of least resistance. The remediation is to drill a stance-frame inventory (In my view / As I see it / My sense is) so the opening frame varies.

The four range-expansion tactics

Tactic 1 — Retrieval-cue chunking

Store vocabulary not as isolated words but as retrieval-ready chunks anchored to functions — a chunk for conceding, a chunk for quantifying, a chunk for hedging. Chunked storage cuts retrieval latency because the candidate retrieves a ready phrase rather than assembling one under load.

Tactic 2 — Semantic-field pre-loading

Before a practice response, pre-load the two or three semantic fields the prompt will require — for a workplace prompt, the fields of scheduling, delegation, and performance. Pre-loading primes retrieval so the relevant vocabulary surfaces faster during the response.

Tactic 3 — Repetition self-monitoring

During practice, the candidate records responses and counts the top three content-word frequencies, then re-delivers the same response substituting a synonym on every second repetition. Self-monitoring builds the internal alarm that fires when a word is about to recur, a discipline the self-monitoring loops and real-time error correction guide treats in depth.

Tactic 4 — Active-vocabulary conversion drills

Move words from passive recognition to active production by drilling them in output, not review. A word recognized on a flashcard is not retrievable in a 90-second response; only a word produced under timed conditions crosses the passive-to-active threshold.

The six-week range-expansion routine

Weeks 1-2 — Repetition baseline and core-verb substitution. Record ten timed responses, compute the type-token ratio for each, and identify the top three recycled content words. Drill the delexicalized-verb matrix and the referential-variation habit. Target: raise the type-token ratio by 0.05.

Weeks 3-4 — Connective and intensifier variation. Layer in the connective inventory and the intensifier ladder. Measure connective diversity (distinct connectives per response) and intensifier diversity; target at least four distinct connectives and three distinct intensifiers per 90-second response.

Weeks 5-6 — Chunk retrieval and semantic pre-loading. Drill retrieval-cue chunks and pre-loading before each practice prompt. Measure retrieval latency indirectly through silent-gap count; target zero gaps above one second while maintaining the raised type-token ratio.

Across all six weeks, the principle is constant: the extended-response tasks do not test how many words the candidate knows — they test how many the candidate can retrieve and deploy in real time. Repetition is a retrieval symptom, not a knowledge gap. A candidate who builds retrieval-ready chunks, pre-loads the right semantic fields, and self-monitors for recurrence converts a broad passive vocabulary into visible spoken range — and moves the language-use band without opening a single new word list. For the pacing discipline that keeps range expansion from slowing delivery, see the response timing and fluency pacing guide.