gre-vocabulary-builder

GRE Vocabulary Builder

The GRE tests vocabulary in one narrow way: can you judge whether a word fits a determinate slot in an academic sentence. That means recognition + nuance discrimination, not production, spelling, or definition recital. Everything here optimizes for that use: words are learned in sentence context, in meaning clusters, with charge attached — because that is exactly the information TC/SE questions interrogate.

Use this when

• The user wants to learn new GRE words or work through a word list.
• The user keeps forgetting specific words and wants a retention routine.
• The user met unknown words in TC/SE practice (handed over from gre-verbal-reasoning).
• The user wants drills: quizzes, cloze sentences, charge-sorting, cluster discrimination.

Do not use this when

• The user is solving TC/SE questions — use gre-verbal-reasoning (it routes vocabulary gaps back here).
• The user needs IELTS topic vocabulary or academic collocations for essays — use ielts-vocabulary-builder.

---

What GRE vocabulary actually is

1. High-frequency test words — the recycled core (~800–1200 words) of academic/literary vocabulary: laconic, prosaic, equivocate, venerate, intransigent… Frequency-ranked lists exist in every major prep resource; working through a high-frequency deck is genuinely worth it, unlike memorizing a 3,500-word dictionary.
2. Secondary meanings of ordinary words — the highest-value, most-neglected category. The GRE loves the meaning you don't know of a word you do: qualify (= limit), check (= restrain), flag (= weaken), appropriate (v., = seize), wanting (= deficient), economy (= efficiency of expression), arrest (= stop), telling (= revealing). When one appears anywhere in practice, capture it explicitly — the familiar spelling actively blocks relearning.
3. Charge and register — many TC/SE eliminations run on polarity alone (is this word approving or critical?). Every word entry must carry its charge.

The word entry format

Never record a bare word–definition pair; record what the test will ask about:

**equivocate** (v.) — to use deliberately ambiguous language to avoid committing
charge: negative (evasive)
cluster: hedge, prevaricate, temporize | ≠ lie outright (that's mendacity)
GRE-style: Pressed for a timeline, the minister ___ed, offering answers
          notable chiefly for their strategic vagueness.
hook: equi-voc — "equal voices," saying two things at once

The cluster line does discrimination work: GRE answer choices routinely put near-neighbors side by side, so learn hedge vs. prevaricate vs. equivocate together, with the distinguishing edge stated. The hook (root, image, or mnemonic — whatever sticks) is memory aid, not etymology lesson; roots (ben-, mal-, -loq-, -phil-, e-/ex-) are recognition boosters for unseen words, not a primary learning method.

Learning protocol

• Batch size: 10–20 new words per session, always introduced in sentences first — give the GRE-style sentence, let the user infer, then confirm. Inference-then-confirmation encodes far better than definition-first.
• Cluster by meaning, not alphabet: teach frugal / parsimonious / penurious / thrifty (and their charge differences) as one unit; alphabetical lists scatter the discriminations the test asks for.
• Spaced retrieval: re-probe new words at roughly 1 day → 3 days → 1 week → 3 weeks, retiring words after two clean, spaced hits in different cue formats. Always test retrieval (blank first), never recognition-by-rereading.
• Leeches (words failed 3+ times): stop repeating the same card. Change the encoding — new hook, new cluster contrast, or a self-written sentence about the user's own research/life. Repetition of a failing encoding just rehearses the failure.

Drill formats (rotate; vary the cue)

1. Cloze (primary) — a GRE-register sentence with a blank, 4–6 choices drawn from the user's own studied words including near-neighbors from the same cluster. This is the closest proxy to the real task.
2. Charge sort — 10 words, sort into positive/negative/neutral in 60 seconds. Fast, high-yield.
3. Cluster discrimination — two near-synonyms, one sentence where only one fits; the user must say why the other fails.
4. Odd one out — 4 words, 3 share a cluster.
5. Reverse cloze — the user writes a sentence in which the target word is forced (a road-sign structure that only that meaning satisfies). Production is not tested on the GRE, but it is the strongest encoding drill.

Vary cue format across re-probes of the same word — a word only ever tested by the same cloze sentence has memorized the sentence, not the word.

Session ledger

Track per word: date introduced, probe history (format + result), status (learning / review / retired / leech). If the user has a practice vault, keep this as a vocab-ledger file there and open sessions by pulling due re-probes before introducing new words — due reviews outrank new material.

Workflow

• New-words session: pull due re-probes → introduce the new batch in sentence context → immediate mixed drill (formats 1–2) covering new + due words → update ledger → end with the 3 shakiest words named.
• "Quiz me": mixed drill from the ledger (or a user-supplied list), weighted toward due and leech words; report accuracy by cluster and any word due for encoding change.
• Word lookup ("what does X mean"): give the full entry format above — including how the GRE tends to use it — not a dictionary line. Offer to add it to the ledger.
• Handoffs from TC/SE practice: words arriving from gre-verbal-reasoning error logs enter as high-priority; they have proven test relevance for this user.