TOEIC Link Grammar — Raising and Control Verb Recognition: How the Subject-Raising, Object-Raising, Subject-Control, and Object-Control Distinctions Discriminate the Upper-Band Grammar Questions

Raising and control verb constructions are the most under-recognized argument-structure distinction in the TOEIC Link grammar segment. On the surface, the two construction families look syntactically identical — a matrix verb takes a noun-phrase argument followed by a non-finite complement clause — but the underlying argument structures are fundamentally different, and the grammar segment's upper-band questions deploy the distinction as a recognition discriminator that the candidate's surface-syntactic reading cannot resolve. Internal practice-corpus data indicates that candidates in the 17-to-20 band recognize the raising-versus-control distinction in roughly two out of ten test items, while candidates in the 25-to-28 band recognize it in eight out of ten items. The recognition gap is the dominant predictor of upper-band grammar segment performance on this question type, and the gap is closable through explicit taxonomy work, diagnostic-test internalization, and timed-condition rehearsal.

The TOEIC Link grammar segment tests the four-way raising-versus-control distinction across recognition questions, sentence-completion questions, and error-detection questions, and each question type rewards the candidate who has produced an explicit argument-structure reading rather than a surface-syntactic reading. For related coverage of the syntactic constructions raising and control verb recognition coordinates with, see the grammar non-finite clause and reduced relative recognition guide, the grammar gerunds and infinitives guide, and the grammar noun clauses and reported speech guide.

The four-way taxonomy

The raising-and-control construction family decomposes into four argument-structure types — subject-raising, object-raising, subject-control, and object-control — and each type instantiates a distinct relationship between the matrix verb's argument structure and the embedded clause's argument structure. The recognition discipline depends on the candidate's explicit taxonomy knowledge.

Subject-raising

Subject-raising verbs do not assign a thematic role to their syntactic subject. The syntactic subject is moved (raised) from the embedded clause's subject position, and the matrix verb contributes only an evidential, modal, or aspectual layer to the proposition the embedded clause expresses.

Representative subject-raising verbs: seem, appear, happen, turn out, tend, be likely, be certain, be sure (in epistemic readings).

Diagnostic property 1: the matrix subject can be an expletive (it, there) when the embedded clause's grammatical structure licenses one. It seems that the report is late. There appears to be a problem. The expletive licensing is impossible with control verbs.

Diagnostic property 2: the matrix subject can be semantically incompatible with the matrix verb in isolation but compatible with the embedded predicate. The earnings tend to fluctuate. (Earnings cannot themselves "tend" — the tendency is over the embedded fluctuation predicate.) The compatibility shift is impossible with control verbs.

Diagnostic property 3: passivization of the embedded clause preserves truth-conditional equivalence. The report seems to have been filed by the auditor ≡ The auditor seems to have filed the report. The equivalence is preserved because the matrix subject does not carry an independent thematic role. The equivalence is broken with control verbs.

Object-raising (exceptional case marking)

Object-raising verbs — also called exceptional-case-marking or ECM verbs — do not assign a thematic role to their syntactic object. The syntactic object is the embedded clause's subject and receives accusative case from the matrix verb, but the thematic role is assigned by the embedded predicate.

Representative object-raising verbs: believe, consider, expect, find, judge, prove, show, suppose, think (with infinitive complement).

Diagnostic property 1: the matrix object can be an expletive when the embedded clause licenses one. We believe there to be a problem. We expect it to be obvious. The expletive licensing is impossible with object-control verbs.

Diagnostic property 2: the matrix object can be semantically incompatible with the matrix verb in isolation. We expect the deadline to be missed. (Deadlines are not themselves "expected by us" in the matrix-object thematic sense — the expectation is over the embedded predicate.)

Diagnostic property 3: the embedded clause can be passivized without truth-conditional shift. We expect the auditor to file the report ≡ We expect the report to be filed by the auditor.

Subject-control

Subject-control verbs assign a thematic role to their syntactic subject, and the subject controls (binds) the implicit subject of the embedded non-finite clause. The matrix subject carries an independent thematic role — typically agent, experiencer, or possessor — and the embedded predicate's subject is referentially identical to the matrix subject.

Representative subject-control verbs: try, attempt, want, hope, plan, decide, refuse, promise, intend, manage, agree, offer.

Diagnostic property 1: the matrix subject cannot be an expletive. *It tries to rain. *There hopes to be a solution. The expletive impossibility is the strongest diagnostic for subject-control versus subject-raising.

Diagnostic property 2: the matrix subject is thematically constrained by the matrix verb. The candidate hopes to pass. (The candidate is the experiencer of hope and the agent of pass — both roles are present.) The thematic doubling is the structural signature of subject-control.

Diagnostic property 3: passivization of the embedded clause breaks truth-conditional equivalence. The auditor tries to file the report ≢ The report tries to be filed by the auditor. The equivalence breaks because the matrix subject carries an independent thematic role that the passive cannot preserve.

Object-control

Object-control verbs assign thematic roles to both their syntactic subject and their syntactic object, and the object controls the implicit subject of the embedded non-finite clause. The matrix verb relates the matrix subject (typically agent or causer) to the matrix object (typically goal, recipient, or theme) and predicates the embedded clause of the matrix object.

Representative object-control verbs: persuade, convince, order, command, force, urge, encourage, allow, permit, advise, ask, tell, instruct, require.

Diagnostic property 1: the matrix object cannot be an expletive. *We persuade it to rain. *We force there to be a solution. The expletive impossibility distinguishes object-control from object-raising.

Diagnostic property 2: the matrix object is thematically constrained by the matrix verb. We persuaded the auditor to file the report. (The auditor is both the persuadee — the goal of persuasion — and the agent of filing.) The thematic doubling is the structural signature of object-control.

Diagnostic property 3: passivization of the embedded clause breaks truth-conditional equivalence. We persuaded the auditor to file the report ≢ We persuaded the report to be filed by the auditor. The second sentence is degraded or anomalous because reports cannot be persuaded.

The diagnostic-test protocol

The diagnostic-test protocol is the candidate's procedural method for disambiguating a target construction across the four-way taxonomy. The protocol applies three diagnostic tests sequentially — the expletive-substitution test, the thematic-compatibility test, and the passivization-equivalence test — and the test pattern identifies the construction type.

Test 1 — Expletive substitution

The expletive-substitution test substitutes an expletive (it, there) for the relevant argument (matrix subject for the subject-raising-versus-subject-control diagnosis, matrix object for the object-raising-versus-object-control diagnosis) and evaluates grammaticality.

Expletive grammatical → raising construction. Expletive ungrammatical → control construction.

The expletive-substitution test is the strongest single diagnostic and produces an immediate raising-versus-control classification when applied correctly. The candidate's failure mode is the application of expletive it in a meaning-shift context where the expletive reading is impossible to evaluate; the candidate must construct the expletive substitution under a meaning-neutral interpretation.

Test 2 — Thematic compatibility

The thematic-compatibility test substitutes a semantically incompatible noun phrase (typically an inanimate object for an animate-required argument or a non-volitional entity for a volition-required argument) for the relevant argument and evaluates whether the matrix verb's thematic constraints are violated.

No violation → raising construction (the argument is not thematically constrained by the matrix verb). Violation → control construction (the argument is thematically constrained).

The thematic-compatibility test is the second diagnostic and produces classification confidence when the expletive test is inconclusive. The candidate's discipline is the construction of a clean thematic-violation context that isolates the matrix verb's thematic constraint from the embedded predicate's constraint.

Test 3 — Passivization equivalence

The passivization-equivalence test passivizes the embedded clause and evaluates truth-conditional equivalence with the active version.

Equivalence preserved → raising construction. Equivalence broken → control construction.

The passivization-equivalence test is the third diagnostic and produces the strongest classification confidence on subtle cases where the expletive and thematic tests are ambiguous. The candidate's discipline is the construction of a clean passive that does not introduce alternative readings beyond the active-passive alternation itself.

The recognition protocol under timed conditions

The recognition protocol the candidate applies under the segment's timed conditions compresses the three-test diagnostic into a two-phase decision procedure that produces the classification within the per-item time budget.

Phase 1 — Surface-pattern classification

The phase-1 work classifies the construction by its surface verb-and-complement pattern. The candidate identifies the matrix verb, the matrix complement type (NP-to-VP, NP-VP, that-clause, NP-that-clause), and the embedded clause's non-finite form (infinitive, gerund, bare infinitive, participle).

The phase-1 output is a verb-and-pattern classification that maps to a candidate set of taxonomy types. Matrix verbs like seem, appear, happen with NP-to-VP complement map to subject-raising or subject-control candidate set; the subsequent phase-2 work disambiguates within the set.

Phase 2 — Diagnostic application

The phase-2 work applies the diagnostic tests sequentially against the candidate set produced in phase 1. The candidate runs the expletive-substitution test first; if conclusive, the classification is complete. If the expletive test is inconclusive, the candidate runs the thematic-compatibility test. If both tests are inconclusive, the candidate runs the passivization-equivalence test.

The phase-2 sequential application is the time-efficient diagnostic procedure under the segment's timed conditions. The candidate's discipline is the disciplined sequence — the expletive test is applied first because it is the fastest test to execute and the most diagnostic test for the most common construction types.

The distractor architecture

The grammar segment's upper-band questions on raising and control constructions deploy four distractor types — the surface-pattern matching distractor, the expletive-licensing inverted distractor, the thematic-role inverted distractor, and the passivization-broken distractor — and each distractor is specifically engineered against a candidate's recognition failure mode.

The surface-pattern matching distractor produces an answer option that matches the question's surface-syntactic pattern but instantiates the wrong raising-or-control type. The candidate whose recognition is surface-syntactic alone selects this distractor because the surface match looks correct. The diagnostic protocol's argument-structure reading rejects this distractor.

The expletive-licensing inverted distractor produces an answer option that licenses or fails to license an expletive in the inverted way for the target construction. The candidate whose phase-2 work skips the expletive test selects this distractor when the surface pattern is shared with the correct answer. The diagnostic protocol's expletive test rejects this distractor.

The thematic-role inverted distractor produces an answer option that assigns thematic roles incompatibly with the target construction's argument structure. The candidate whose phase-1 work does not classify the matrix verb's argument-structure properties selects this distractor. The diagnostic protocol's thematic-compatibility test rejects this distractor.

The passivization-broken distractor produces an answer option whose passive version is truth-conditionally inequivalent to the active version in the way characteristic of control rather than raising (or vice versa). The candidate whose phase-2 work does not apply the passivization test selects this distractor. The diagnostic protocol's passivization-equivalence test rejects this distractor.

The rehearsal sequence

The raising-and-control rehearsal sequence produces band-stable recognition competence across the four-phase preparation cycle the segment's upper-band content rewards.

The taxonomy-consolidation phase produces the candidate's explicit knowledge of the four construction types and the representative verb classes each instantiates. The candidate works through the taxonomy with verb-class-organized practice and produces the per-type recognition the diagnostic protocol depends on.

The diagnostic-protocol internalization phase produces the candidate's three-test diagnostic execution under non-timed conditions. The candidate works through ambiguous-construction examples with explicit phase-1 surface classification and phase-2 diagnostic application and produces the classification competence the segment's upper-band questions reward.

The distractor-architecture calibration phase produces the candidate's distractor-detection competence across the four distractor types. The candidate works through grammar-segment practice items at the upper-band level with explicit distractor-type identification and produces the distractor-rejection competence the segment's scoring rewards.

The timed-condition consolidation phase produces the candidate's section-time-stable recognition and production. The candidate works through raising-and-control content under the segment's per-item time budget (typically thirty-five to forty-five seconds for grammar-segment items at the upper-band level) and produces the timing-stable competence the live segment requires.

The rehearsal-sequence completion produces the band-stable raising-and-control recognition competence the segment's upper-band grammar questions require and supports the candidate's upper-band scoring on the argument-structure-themed content the segment increasingly deploys at the 21-and-above bands.

For related coverage of the grammar disciplines that raising-and-control competence coordinates with, see the grammar passive voice and causative guide and the grammar subjunctive mood and hypothetical constructions guide.