H1 General Paper Paper 2 Connections questions test whether a candidate can hold more than one writer’s position accurately at the same time. The challenge is not simply to find a similar phrase in two passages. It is to identify the exact relationship between propositions: agreement on cause but disagreement on severity; agreement on the problem but disagreement on responsibility; a shared conclusion reached through different values; or one writer qualifying what another states more absolutely.
This page is the SETC canonical owner for the Paper 2 cross-passage Connections examination-performance layer. It does not replace the broader English Source Synthesis Performance owner, which covers transferable multi-text reading and writing. This page owns the narrower Paper 2 performance sequence: preserve passage identity, abstract propositions, label the relationship, locate supporting lines, express the connection economically and move on within the 90-minute paper.
Under the current 2027 H1 GP syllabus 8881, Paper 2 includes 4–6 marks of connections of ideas across Passages 1 and 3 and Passages 2 and 3. Future candidates should verify the syllabus for their own examination year with SEAB before relying on exact mark or format details.
Connections are relationships between claims, not topics
Two writers can discuss the same topic and disagree completely. Two writers can use different topics and still make structurally similar claims. Surface overlap is therefore a weak starting point. The candidate must abstract what each writer is actually saying.
“Both passages discuss artificial intelligence” is not a connection. “Both writers argue that AI can widen access, but Passage 1 treats lower cost as the main mechanism while Passage 3 emphasises personalised support” is a connection because the relationship is explicit.
The Connections operating model
- Preserve passage identity. Know who argues what.
- Abstract the proposition. Restate each relevant claim in plain conceptual language.
- Choose the comparison dimension. Cause, consequence, severity, responsibility, solution, value, evidence, condition or degree.
- Label the relationship. Agree, disagree, qualify, extend, narrow, reverse, complement or shift emphasis.
- Locate textual support. Select the lines that actually express the proposition.
- Write the relationship first. Do not merely alternate quotations.
- Preserve important difference. Similarity should not erase nuance.
- Control scope and quantity. Answer exactly the required connection.
- Close efficiently. Protect time for Application.
Iona dominates Connections
Iona must preserve source attribution and claim scope. Her main risk is false equivalence: deciding that two passages agree because they contain similar vocabulary, or deciding that they disagree because their examples look different.
Maren helps organise the answer. A Connections response should not become two mini-summaries. The relationship should lead, and the evidence should support it.
Leonie protects time. Cross-passage evidence can be far apart; a candidate needs location memory and a bounded search rather than rereading entire passages repeatedly.
Passage identity is non-negotiable
Before Connections work, the learner should be able to state one line for each passage: central claim, main reasoning route and key qualification. If the writer’s final position cannot be stated accurately, comparison is premature.
Abstract before comparing
Suppose Passage 1 discusses declining trust in newspapers and Passage 3 discusses distrust of government agencies. The surface subjects differ. At a higher level, both may argue that opacity weakens institutional legitimacy. The candidate must decide whether that abstraction is faithful and relevant to the question.
Abstraction should be high enough to reveal a relationship but not so high that all differences disappear. “Both discuss trust” may be too vague. “Both argue that perceived lack of transparency reduces trust” may be appropriately precise.
Choose a comparison dimension
Writers may be compared on cause, consequence, seriousness, responsibility, solution, moral value, feasibility, evidence, time horizon, stakeholder or condition. Many weak responses fail because they compare different dimensions without noticing.
For example, Passage 1 may explain why misinformation spreads, while Passage 3 explains who should respond. These are related but not directly contradictory. A candidate should not manufacture disagreement across different questions.
Agreement can be partial
Two writers may agree on the existence of a problem but disagree on its magnitude. They may agree on the goal but disagree on the method. They may agree on a cause but disagree on whether it can be solved. State the exact level of agreement rather than writing “both agree” without qualification.
Disagreement can be narrow
Two writers may reach similar conclusions but differ on why. They may both support regulation, while one grounds the case in fairness and the other in public safety. That difference matters because it affects how each would respond to exceptions.
Qualification is a relationship
Passage 3 may not reject Passage 1; it may limit the circumstances in which Passage 1’s claim holds. “Passage 3 qualifies Passage 1 by arguing that the benefit appears only where institutions are strong” is more precise than “Passage 3 disagrees”.
Extension is a relationship
One passage may take another idea further. Passage 1 argues that technology changes work; Passage 3 extends this by arguing that it changes political power through control of data. The second passage adds consequence rather than contradicting the first.
Complement is a relationship
Two passages can explain different parts of the same mechanism. One identifies an individual behaviour; another identifies institutional incentives that reinforce it. A synthesis answer can show how the explanations complement each other.
Shift in responsibility is a relationship
One writer may place responsibility on individuals, another on corporations, and another on governments. The topic is identical, but the causal and moral allocation differs. Connections questions often reward this kind of exact distinction.
Shift in time horizon is a relationship
One passage may focus on immediate benefits while another emphasises long-term costs. This is not necessarily disagreement about facts. It may be a difference in horizon. Naming the horizon clarifies the relationship.
Shift in stakeholder is a relationship
One writer can assess a policy from the perspective of consumers, another from workers, another from the state. Apparent disagreement may arise because the passages optimise different outcomes. A candidate should notice whose interests structure each claim.
Relationship first, evidence second
Before searching for lines, write the relationship in simple form: “Both agree X, but P3 limits X to Y”; “P1 blames A, whereas P3 assigns responsibility to B”; “P3 extends P2 by adding consequence C.” Then locate the lines that genuinely support the statement.
This prevents evidence hunting from determining the interpretation accidentally.
Do not alternate mini-summaries
“Passage 1 says X. Passage 3 says Y.” may be factually accurate but leaves the relationship implicit. A stronger answer uses comparative grammar: while, whereas, both, unlike, qualifies, extends, similarly, in contrast—only where the relationship is accurate.
Do not force a binary
Not every relationship is agree/disagree. If the passages operate at different levels, use language such as extends, complements, shifts emphasis, narrows, broadens or applies the same principle differently.
Do not flatten strength
If Passage 1 says regulation is essential and Passage 3 says it may be useful under some conditions, the passages are not identical. Preserve modal strength when describing the connection.
Do not flatten values
Two writers may support the same policy for different reasons. One values freedom, another equality. The shared recommendation does not erase the different argumentative foundations.
Do not flatten mechanism
Two passages may predict the same outcome through different causal pathways. State the pathway if it is relevant. This can be more insightful than merely noting agreement on outcome.
Use proposition pairs
During training, write one sentence for each relevant passage claim before comparing. Example: P1—“Government regulation reduces platform harm by changing incentives.” P3—“User education is more effective because harmful behaviour adapts around regulation.” Now the disagreement is visible: mechanism and preferred intervention.
Use connection matrices
Create a table with dimensions such as cause, effect, responsibility, solution, severity and value. Fill only where the passage has a clear position. This reveals where direct comparison is legitimate and where passages simply address different aspects.
Use relationship labels as temporary scaffolds
Agreement, disagreement, qualification, extension, complement, shift of emphasis, different cause, different solution, different value and different stakeholder are useful labels during learning. They should eventually become natural comparative language rather than formulaic headings.
Evidence may be non-adjacent
A writer’s position may be established in one paragraph and qualified later. Do not compare an early absolute-sounding sentence if the final stance is narrower. Passage mapping should guide evidence selection.
Source attribution must remain exact
With three passages, it is easy to remember the right idea and attach it to the wrong writer. Check attribution before writing. A conceptually correct relationship can still fail if the positions are reversed.
Synthesis should preserve difference
The purpose of synthesis is not to make the writers sound the same. A strong synthesis can state a shared premise and then expose the decisive difference. “Both writers accept that misinformation threatens trust, but Passage 1 sees platform incentives as the root problem whereas Passage 3 sees citizens’ weak media literacy as more fundamental.”
Connection answers should be economical
The task is usually narrow. Avoid reproducing entire passage summaries. State relationship, support from each passage and the relevant distinction. Then close. Paper 2 time must still serve Application.
Connections Diagnostic Casebook | Sixty Failure Patterns and Repairs
Case 1: Same topic mistaken for agreement
Repair compares propositions, not subject matter.
Case 2: Different examples mistaken for disagreement
Repair abstracts the underlying claims.
Case 3: Agreement on problem, disagreement on solution
Repair names both levels explicitly.
Case 4: Agreement on solution, disagreement on cause
Repair compares mechanism separately from recommendation.
Case 5: Disagreement only in degree
Repair preserves modal strength.
Case 6: Qualification labelled contradiction
Repair distinguishes limiting conditions from reversal.
Case 7: Extension labelled agreement only
Repair states what new consequence or dimension is added.
Case 8: Complement labelled disagreement
Repair asks whether both explanations can be true simultaneously.
Case 9: Same policy, different values
Repair identifies moral foundation rather than recommendation alone.
Case 10: Same outcome, different mechanisms
Repair compares causal routes.
Case 11: Same cause, different predicted outcomes
Repair identifies consequence as comparison dimension.
Case 12: Different stakeholder focus
Repair names whose interests structure each argument.
Case 13: Different time horizon
Repair distinguishes short-term and long-term assessments.
Case 14: Different scale
Repair notes individual versus societal, local versus global or pilot versus system-level claim.
Case 15: Similar wording, opposite meaning
Repair reads proposition in context rather than keyword match.
Case 16: Different wording, same proposition
Repair abstracts meaning before comparison.
Case 17: Early Passage 1 claim compared to final Passage 3 claim
Repair updates each passage identity to final qualified stance.
Case 18: Candidate reverses passage attribution
Repair uses identity grid and source check.
Case 19: Candidate quotes correct lines but no relationship
Repair writes relation first.
Case 20: Candidate states relationship with no evidence
Repair adds precise support from both passages.
Case 21: Candidate summarises P1 then P3
Repair integrates comparative grammar.
Case 22: Candidate uses “both agree” without dimension
Repair states what they agree about.
Case 23: Candidate uses “they disagree” without dimension
Repair states cause, solution, value or degree.
Case 24: Candidate overstates disagreement
Repair compares modal and evaluative strength.
Case 25: Candidate overstates agreement
Repair preserves the qualification that separates positions.
Case 26: Candidate compares evidence instead of claims when question asks ideas
Repair answers the required level.
Case 27: Candidate compares claims when question asks reasons
Repair aligns comparison dimension to command.
Case 28: Candidate ignores shared assumption
Repair identifies common underlying premise when relevant.
Case 29: Candidate ignores different assumption
Repair explains why similar facts lead to different conclusions.
Case 30: Candidate compares tone rather than idea
Repair distinguishes stylistic relation from conceptual relation unless tone is the target.
Case 31: Candidate compares examples instead of mechanisms
Repair abstracts what each example is used to show.
Case 32: Candidate treats complement as redundancy
Repair shows how two explanations operate at different levels.
Case 33: Candidate treats extension as contradiction
Repair asks whether the added claim can coexist with the first.
Case 34: Candidate treats different solution as different problem
Repair separates diagnosis from remedy.
Case 35: Candidate ignores writer’s concession
Repair includes qualified agreement where it affects relationship.
Case 36: Candidate treats concession as final stance
Repair tracks what follows the concession.
Case 37: Candidate cannot find evidence quickly
Repair improves passage-location memory and proposition map.
Case 38: Candidate rereads whole passages repeatedly
Repair uses identity grid and structural map.
Case 39: Candidate spends too long making answer elegant
Leonie applies sufficiency and moves on.
Case 40: Candidate writes relationship too vaguely
Repair names exact proposition and dimension.
Case 41: Candidate writes too much passage detail
Repair selects only evidence necessary to establish relationship.
Case 42: Candidate leaves relationship implicit
Repair uses explicit comparative clause.
Case 43: Candidate uses stock “similarly” incorrectly
Repair checks whether propositions actually align.
Case 44: Candidate uses stock “in contrast” incorrectly
Repair identifies whether difference is true opposition or merely emphasis.
Case 45: Candidate cannot compare abstract passages
Repair reduces each to claim, mechanism, value and condition.
Case 46: Candidate relies on background knowledge
Repair compares only what passages actually claim.
Case 47: Candidate sees relation but attributes wrong reason
Repair rechecks each writer’s causal logic.
Case 48: Candidate treats one passage example as writer’s central position
Repair distinguishes evidence from thesis.
Case 49: Candidate ignores passage qualification in connection answer
Repair preserves final stance.
Case 50: Candidate comparison is correct but source lines weak
Repair chooses more direct evidence.
Case 51: Candidate evidence is strong but comparison wrong
Repair returns to proposition relationship before quoting.
Case 52: Candidate memorises comparison phrases
Repair generates wording from actual relation.
Case 53: Candidate merges two writers into one voice
Repair maintains passage labels during planning.
Case 54: Candidate compares only opening paragraphs
Repair uses global stance, including later qualification.
Case 55: Candidate loses time searching one elusive connection
Repair uses bounded effort and protects Application time.
Case 56: Candidate skips connection work too early
Repair requires one relationship hypothesis before deferral.
Case 57: Candidate changes correct relation during review
Repair requires new textual evidence before changing.
Case 58: Candidate strong only when passages strongly oppose
Repair trains subtle qualification, extension and partial agreement.
Case 59: Candidate strong only when vocabulary overlaps
Repair uses paraphrased proposition matching.
Case 60: Candidate panics when connection is not obvious
Repair runs the matrix: cause, effect, responsibility, solution, severity, value, condition. One dimension often reveals the relation.
Six Connections training cycles
Cycle 1: Passage identity. Summarise claims and qualifications before comparison.
Cycle 2: Proposition abstraction. Restate surface-different claims at comparable conceptual level.
Cycle 3: Relationship labels. Train agreement, disagreement, qualification, extension, complement and shift.
Cycle 4: Evidence matching. Write relationship first, then locate direct support from each passage.
Cycle 5: Synthesis writing. Express relationship in one integrated comparative structure.
Cycle 6: Timed integration. Complete Connections after earlier Paper 2 tasks and protect Application time.
Connections tutor audit
- Can the learner state each passage’s final position accurately?
- Can surface wording be abstracted into propositions?
- Does the learner choose one comparison dimension at a time?
- Can partial agreement and limited disagreement be expressed?
- Can qualification, extension and complement be distinguished from contradiction?
- Are source attributions accurate?
- Does relationship come before evidence?
- Are important modal and value differences preserved?
- Can synthesis integrate rather than alternate passage summaries?
- Are answers economical enough to protect Paper 2 time?
Connections examination-day card
- Restate each relevant passage claim.
- Choose the comparison dimension.
- Label the relationship precisely.
- Check whether agreement/disagreement is full, partial or conditional.
- Locate the strongest supporting line in each passage.
- Write the relationship first.
- Preserve difference in strength, value, cause or context.
- Avoid two separate mini-summaries.
- Close once the relationship is proven.
- Protect time for Application.
The Connections standard
World-class Connections performance means the candidate can hold several writers’ ideas accurately without flattening them into “same” or “different”. They identify the precise dimension of comparison, preserve qualifications and causal differences, support the relationship from both texts and express the synthesis economically under time.
Properly taught kids shine a bright light into the future. In Connections, that light is comparative precision: see what the writers truly share, see where they genuinely diverge, and do not erase the difference just to make the answer easier.
Where to go next
- Return to H1 GP Paper 2 Examination Performance.
- Use Passage 1, Summary and Application child guides for their distinct task systems.
- Use English Source Synthesis Performance for the broader transferable multi-text mechanism.