PHASE 4: CROSS-REFERENCE ANALYSIS

BR Clusters (63) vs. Reference Vocabulary (47 codes)

Taxonomy Extraction Project | Processed: 2026-02-19

SECTION 1: MAPPING TABLE

SECTION 2: GAPS IN REFERENCE

The following BR clusters have no equivalent in the 47-code reference vocabulary.

GAP 1: C13 — DEMONSTRATIVE_PROOF

• Function: Visual or tangible demonstration of the product/method in action (before/after photos, live demonstrations, product application videos). Shows rather than tells.
• Why the reference misses it: The reference vocabulary categorizes proof by social dimension (PROOF_SOCIAL, PROOF_AUTHORITY, PROOF_SPECIFICITY) but not by modality. Visual demonstration is a distinct proof type that carries unique persuasive weight in video-based DR formats. Written-copy frameworks may underweight this because it requires a visual medium.
• Proposed code: PROOF_DEMONSTRATION — visual or tangible demonstration of the product/method working in real-time or via before/after evidence.

GAP 2: C30 — PRICE_BENEFIT

• Function: Positively framing the product's affordability, value, and cost advantage vs. alternatives. Includes price anchoring ("worth $500, yours for $49"), daily cost reframing ("less than a cup of coffee"), and comparative value claims.
• Why the reference misses it: The reference has INVALIDATE_ON_PRICE (attacking competitors' prices) but not the positive mirror — promoting your own price as a benefit. These are opposite persuasive vectors. Attacking competitor pricing is an invalidation move; promoting your own value is an offer-building move.
• Proposed code: VALUE_FRAME — positively framing the product's price as a benefit through anchoring, comparison, or reframing.

GAP 3: C34 — SCIENTIFIC_DISCOVERY

• Function: Framing the solution's origin as a recent breakthrough scientific or medical discovery. Wraps credibility in a narrative of novelty ("scientists just discovered...," "a new study from Harvard reveals...").
• Why the reference misses it: The reference has proof codes (PROOF_AUTHORITY, PROOF_SPECIFICITY) and a curiosity code (CURIOSITY_OPEN), but no code for the specific narrative device of "discovery framing" that combines novelty, authority, and curiosity simultaneously. It is a hybrid device, which may explain why it fell through categorical cracks.
• Proposed code: DISCOVERY_FRAME — framing the solution's origin as a recent, breakthrough scientific or medical discovery to combine novelty-curiosity with institutional authority.

GAP 4: C41 — STORY_EMOTIONAL

• Function: Using personal narrative, emotional stories, or biographical elements to create empathetic connection. The copywriter tells a story (their own or a character's) to transport the viewer emotionally before delivering the persuasive payload.
• Why the reference misses it: The reference vocabulary is organized by persuasive function (what the copy does) rather than narrative device (how the copy delivers it). Storytelling is a delivery vehicle that can carry pain articulation, proof, hope bridging, and other functions simultaneously, making it hard to place in a function-only taxonomy.
• Proposed code: STORY_TRANSPORT — personal or character-driven narrative used to create empathetic connection and transport the viewer emotionally.

GAP 5: C42 — THE_ONE_THING

• Function: Positioning the entire message around a single, crucial insight or action that changes everything. Creates a focusing frame: "there is ONE thing you need to know/do."
• Why the reference misses it: The reference has CURIOSITY_OPEN (opens loops) and ROOT_CAUSE_REVEAL (names the cause), but THE_ONE_THING is a framing/positioning device that promises singularity and simplicity before the reveal. It is a pre-reveal frame, not the reveal itself.
• Proposed code: SINGULAR_INSIGHT — framing the message around a single, crucial insight or action that changes everything.

GAP 6: C44 — PRODUCT_BUILDING

• Function: Narrating the creation journey of the product -- the expert's frustration with existing solutions, their research process, and their breakthrough moment. Functions as a credibility + empathy + differentiation combo.
• Why the reference misses it: The reference has EXPERT_PRESENT (introducing credentials) and MSOL_EXPLAIN (explaining mechanism), but no code for the product's origin story as a narrative arc. This is common in Brazilian DR where the product is presented as the expert's personal quest.
• Proposed code: ORIGIN_STORY — narrating the creation journey of the product through the expert's research, frustration, and breakthrough.

GAP 7: C46 — AUDIENCE_EXPANSION

• Function: Deliberately broadening the target audience by listing additional problems, symptoms, or demographics the method addresses. "This also works for headaches, joint pain, insomnia..."
• Why the reference misses it: The reference has WORKS_FOR_ALL (universal applicability) and SELF_SELECT/QUALIFY (narrowing the audience), but no code for actively expanding the problem set to recruit more viewers. AUDIENCE_EXPANSION widens the funnel; WORKS_FOR_ALL claims universality. They are complementary but distinct.
• Proposed code: AUDIENCE_EXPAND — deliberately broadening the target audience by listing additional problems, symptoms, or demographics addressed.

GAP 8: C49 — BULLETS_LIST

• Function: Listing multiple benefits, features, or problems in a bullet-point or enumerated format. A structural/pacing device that accelerates information delivery.
• Why the reference misses it: The reference vocabulary focuses on persuasive functions, not presentation formats. Bullet lists are a formatting convention, not a unique persuasive move. However, in video DR, the rapid-fire bullet format serves a distinct pacing function (information density + momentum building).
• Proposed code: BULLET_STACK — rapid-fire enumeration of benefits, features, or problems in list format for information density and momentum.

GAP 9: C51 — MARKETING_THESIS

• Function: Explaining the underlying marketing/scientific rationale at a meta-strategic level. The copy breaks the fourth wall to explain why this approach is different at a conceptual level.
• Why the reference misses it: Only 1 occurrence in the BR corpus. Extremely rare. The reference vocabulary does not include meta-level marketing strategy explanations because they are uncommon in consumer DR. This may be an outlier rather than a systematic gap.
• Proposed code: META_RATIONALE — meta-level explanation of the marketing or scientific strategy behind the product positioning.

GAP 10: C52 — ANALOGY

• Function: Using comparison or metaphor to make a complex mechanism understandable. A rhetorical/pedagogical device.
• Why the reference misses it: Like BULLETS_LIST, this is a rhetorical tool rather than a persuasive function. The reference vocabulary does not code for rhetorical devices, only for strategic persuasion moves. Analogies are a teaching technique that makes MUP_EXPLAIN or MSOL_EXPLAIN more effective but are not themselves a distinct persuasive function.
• Proposed code: ANALOGY_DEVICE — comparison or metaphor used to make a complex mechanism understandable.

GAP 11: C57 — GOSSIP_INTIMATE_TONE

• Function: Using an intimate, gossipy, girlfriend-to-girlfriend conversational tone as a persuasion device. Culturally specific to Brazilian female-targeted DR (fofoca).
• Why the reference misses it: The reference vocabulary does not code for tonal/register choices. Tone is typically considered a stylistic layer rather than a structural persuasion move. However, in Brazilian female-targeted DR, the fofoca register is a deliberate strategic choice that builds trust and lowers defenses.
• Proposed code: INTIMATE_TONE — deliberately informal, gossip-register conversational tone used to build trust and lower defenses.

GAP 12: C60 — STEALTH_SELLING

• Function: Disguising the sales pitch as casual conversation, podcast content, or friendly recommendation. The persuasive intent is concealed.
• Why the reference misses it: The reference vocabulary assumes overt persuasion (the viewer knows they are being sold to). Stealth selling is a meta-technique that wraps the entire persuasion structure in a non-commercial frame. This is increasingly common in social media DR where ads are designed to not look like ads.
• Proposed code: STEALTH_FRAME — disguising the sales pitch as non-commercial content (conversation, podcast, friendly recommendation).

GAP 13: C61 — GAP_TO_VSL

• Function: Creating an information gap that drives the viewer from a short-form ad (e.g., Facebook/Instagram video) to the full VSL (Video Sales Letter). A media-bridge function.
• Why the reference misses it: The reference vocabulary assumes a single-format context (one long-form VSL). Brazilian DR commonly uses short-form social media videos as entry points that bridge to full VSLs. The funnel-transition function is specific to multi-step, multi-format campaigns.
• Proposed code: FUNNEL_BRIDGE — information gap or teaser that drives the viewer from the current format to the next step in the funnel (typically short-form ad to full VSL).

SECTION 3: GAPS IN BR

The following reference codes have no BR cluster that maps to them (directly or partially).

GAP_IN_BR 1: SCROLL_STOP (#1)

• Meaning: Pattern interrupt that halts thumb-scrolling. A visual or auditory shock in the first 1-3 seconds designed purely to stop the scroll.
• Possible reasons for absence: The BR annotators focused on persuasive content functions rather than media-specific attention mechanics. SCROLL_STOP is a platform-specific device (social media feeds) that may have been implicitly present in the ads but was not labeled as a distinct annotation. The BR corpus may have categorized scroll-stopping elements under CURIOSITY_HOOK (C31) or BELIEF_DISRUPTION (C33) without isolating the pure attention-capture moment.

GAP_IN_BR 2: SELF_SELECT (#2)

• Meaning: "If you are X, keep watching" qualifier early in the ad.
• Possible reasons for absence: PARTIALLY covered by C24 (AUDIENCE_QUALIFICATION), which includes "E pra voce" and "Qualificador." However, the specific "keep watching" phrasing as a retention device was not isolated as a separate function. The BR annotations merged early-funnel qualification with late-funnel qualification under a single cluster. Listed as GAP because the self-selection-as-retention function is not distinctly captured.

GAP_IN_BR 3: CONSPIRACY_LEAN (#8)

• Meaning: Hinting that the truth is being suppressed by powerful interests.
• Possible reasons for absence: The BR annotations captured COMMON_ENEMY (C20) which names an external villain, and INFORMATION_SCARCITY (C19) which implies suppression ("this video may be taken down"). The conspiracy lean -- the specific insinuation that truth is being hidden -- is distributed across these two clusters rather than isolated. Brazilian DR may rely more on enemy-framing than explicit conspiracy language, or annotators may have subsumed conspiracy hints under the enemy frame.

GAP_IN_BR 4: IDENTITY_SPEAK (#9)

• Meaning: "People like us do / believe X." In-group identity signaling.
• Possible reasons for absence: The BR corpus captured AUDIENCE_QUALIFICATION (C24, who this is for) and EXCLUSIVITY_UNIQUENESS (C59, elite positioning) but did not isolate tribal identity language. Brazilian DR in this corpus may favor qualification over identity reinforcement, or the annotators may not have distinguished identity statements from qualification statements. The function is subtly different: qualification says "this is for you," while identity speak says "we are the kind of people who..."

GAP_IN_BR 5: CREDIBILITY_SEED (#10)

• Meaning: Early, small credibility signal placed before the full authority reveal. A preliminary trust-building micro-moment.
• Possible reasons for absence: The BR annotations captured EXPERT_PRESENTATION (C16), MEDICAL_AUTHORITY (C17), and SCIENTIFIC_PROOF (C14) but did not isolate early-stage credibility signals distinct from full authority reveals. Brazilian DR may front-load authority more aggressively (going straight to expert presentation) rather than using subtle credibility seeds. Alternatively, annotators may have grouped early credibility signals with the full authority beat.

GAP_IN_BR 6: TRANSITION_PIVOT (#22)

• Meaning: Structural beat that changes topic. A generic scene-change or topic-shift marker.
• Possible reasons for absence: The BR annotations focused on what each section does (its persuasive function) rather than the transitions between sections. HOPE_BRIDGE (C40) and CTA_BUILDING (C09) are specific transition types that were captured, but the generic "topic change" beat was not annotated as a distinct function. This makes sense: a transition pivot is a structural marker, not a persuasive function, so content-focused annotators would skip it.

GAP_IN_BR 7: PROOF_SPECIFICITY (#24)

• Meaning: Precise numbers, dates, percentages as proof elements.
• Possible reasons for absence: The BR annotations captured proof by type (social, personal, demonstrative, scientific, testimonial) but not by specificity level. Specific numbers and dates were embedded within other proof annotations (e.g., a scientific proof annotation might contain "87% of users reported improvement") without being called out separately. The BR taxonomy classifies proof by source/type; the reference vocabulary classifies it by evidentiary quality.

GAP_IN_BR 8: PROMISE_TIMELINE (#26)

• Meaning: Result + deadline: "in 7 days you'll..."
• Possible reasons for absence: PARTIALLY covered. C28 (PROMISE) captures result promises broadly, and C58 (RAPID_RESULT) captures speed emphasis. But the specific combination of "result + explicit deadline" was not isolated as a distinct code. The BR annotations treated timeline promises as a subset of general promises or rapid results. Listed as a gap because the explicit deadline component is not distinctly captured.

GAP_IN_BR 9: CTA_EMOTIONAL (#35)

• Meaning: CTA wrapped in an emotional frame.
• Possible reasons for absence: C10 (CTA_COMPOUND) captures CTAs combined with emotional elements, and C56 (MOMENT_OF_TRUTH) captures dramatic CTA transitions. But the specific "emotionally-wrapped CTA" as a distinct code was not isolated. The BR annotations either tagged the CTA as direct (C08) or noted the compound nature (C10) without separately labeling the emotional wrapping.

GAP_IN_BR 10: Emotional_Mobilization (#42)

• Meaning: Stirring enough emotion to trigger action. A meta-function that describes the cumulative emotional intensity needed for conversion.
• Possible reasons for absence: This is a meta-level descriptor of overall emotional intensity rather than a specific copywriting beat. The BR annotations captured specific emotional devices (FEAR_DEEPENING, PAIN_DEEPENING, FUTURE_PACING, etc.) but not the aggregate "enough emotion to act" threshold. This may be too abstract to annotate at the tag level -- it is an emergent property of multiple tags working together.

GAP_IN_BR 11: Tribal_Mobilization (#43)

• Meaning: "Join your tribe" social pull. Inviting the viewer to become part of a group.
• Possible reasons for absence: The BR corpus captured EXCLUSIVITY_UNIQUENESS (C59, elite positioning) and AUDIENCE_QUALIFICATION (C24, who this is for) but not explicit tribal invitation language. Brazilian DR in this health/beauty corpus may not use explicit tribe-building language. This function is more common in identity-driven niches (fitness, political, spiritual) than in health supplement/beauty DR.

GAP_IN_BR 12: Viral_Trigger (#44)

• Meaning: Share-worthy hook or reveal designed to encourage organic sharing.
• Possible reasons for absence: The BR corpus consists of paid direct-response ads, not organic content. Viral triggers are optimized for shareability, which is less relevant when the content is served via paid distribution. The BR annotators may not have observed or labeled share-optimization elements because the ads were designed for conversion, not virality.

GAP_IN_BR 13: Identity_Close (#45)

• Meaning: "You're the kind of person who..." identity framing at the CTA moment.
• Possible reasons for absence: The BR annotations captured CTA_DIRECT (C08) and CTA_BUILDING (C09) but did not isolate identity-based closing language. This specific technique may be less common in Brazilian Portuguese DR, or it may have been subsumed under CTA_COMPOUND (C10) when CTAs were combined with "Seja unico" or qualification elements.

GAP_IN_BR 14: Pause-for-emphasis (#46)

• Meaning: Silence or beat for dramatic effect. A performative/delivery technique.
• Possible reasons for absence: The BR annotations were text-based analysis of bold annotations within a swipe file. Pauses, silence, and delivery beats are audiovisual elements that would not appear in text-based annotation. If the annotator was working from transcripts or written swipe files rather than watching video, pauses would be invisible.

GAP_IN_BR 15: Curiosity_Amplifier (#47)

• Meaning: Loop-within-a-loop that re-opens curiosity mid-ad.
• Possible reasons for absence: The BR annotations captured CURIOSITY_HOOK (C31) at the opening and GAP_TO_VSL (C61) as a funnel bridge, but did not isolate mid-ad curiosity re-openers. The specific "loop within a loop" technique may have been embedded within compound annotations (e.g., "Escassez da informacao + Spoiler do MSOL + Simplicidade + Hook") without being labeled as a distinct curiosity amplification moment. Alternatively, Brazilian DR may rely more on forward-propulsion through new content than on recursive curiosity loops.

SECTION 4: MERGE CANDIDATES

The following BR clusters and reference codes overlap sufficiently to warrant merging under a single unified code.

MERGE 1: C08 CTA_DIRECT + CTA_MECHANICAL (#34)

• Proposed unified code: CTA_DIRECT
• Rationale: Identical function. The BR name is clearer and less ambiguous than "mechanical."

MERGE 2: C09 CTA_BUILDING + CTA_BUILD (#36)

• Proposed unified code: CTA_BUILD
• Rationale: Identical function. Minor naming difference only.

MERGE 3: C11 SOCIAL_PROOF + PROOF_SOCIAL (#25)

• Proposed unified code: PROOF_SOCIAL
• Rationale: Identical function. Adopt reference naming convention (PROOF_ prefix) for consistency with PROOF_AUTHORITY, PROOF_SPECIFICITY.

MERGE 4: C16 EXPERT_PRESENTATION + EXPERT_PRESENT (#41)

• Proposed unified code: EXPERT_PRESENT
• Rationale: Identical function. Adopt shorter reference name.

MERGE 5: C20 COMMON_ENEMY + ENEMY_FRAME (#7)

• Proposed unified code: ENEMY_FRAME
• Rationale: Identical function. ENEMY_FRAME is more descriptive of the technique (framing an enemy) vs. just naming it.

MERGE 6: C21 FEAR_DEEPENING + FEAR_DEEPEN (#4)

• Proposed unified code: FEAR_DEEPEN
• Rationale: Identical function. Adopt the verb form for consistency.

MERGE 7: C23 UNIVERSAL_APPLICABILITY + WORKS_FOR_ALL (#32)

• Proposed unified code: WORKS_FOR_ALL
• Rationale: Identical function. WORKS_FOR_ALL is more intuitive and less formal.

MERGE 8: C25 METHOD_SIMPLICITY + METHOD_SIMPLE (#30)

• Proposed unified code: METHOD_SIMPLE
• Rationale: Identical function. Adopt adjective form for consistency with METHOD_NATURAL.

MERGE 9: C26 FUTURE_PACING + FUTURE_PACE (#27)

• Proposed unified code: FUTURE_PACE
• Rationale: Identical function. Adopt verb form.

MERGE 10: C31 CURIOSITY_HOOK + CURIOSITY_OPEN (#3)

• Proposed unified code: CURIOSITY_OPEN
• Rationale: Identical function. CURIOSITY_OPEN is more precise (the hook opens a loop, not just catches attention).

MERGE 11: C33 BELIEF_DISRUPTION + INVALIDATE_BELIEF (#13)

• Proposed unified code: INVALIDATE_BELIEF
• Rationale: Identical function. INVALIDATE_ prefix groups it with other invalidation codes.

MERGE 12: C35 ROOT_CAUSE_REVEAL + ROOT_CAUSE_REVEAL (#16)

• Proposed unified code: ROOT_CAUSE_REVEAL
• Rationale: Already identical in name and function.

MERGE 13: C36 SUPERSTRUCTURE + SUPERSTRUCTURE (#29)

• Proposed unified code: SUPERSTRUCTURE
• Rationale: Already identical in name and function.

MERGE 14: C38 SKEPTICISM_HANDLING + SKEPTICISM_DISARM (#20)

• Proposed unified code: SKEPTICISM_DISARM
• Rationale: Identical function. DISARM is more active/descriptive than HANDLING.

MERGE 15: C39 REASON_WHY + REASON_WHY (#19)

• Proposed unified code: REASON_WHY
• Rationale: Already identical in name and function.

MERGE 16: C40 HOPE_BRIDGE + HOPE_BRIDGE (#21)

• Proposed unified code: HOPE_BRIDGE
• Rationale: Already identical in name and function.

MERGE 17: C45 DESIRE_ARTICULATION + DESIRE_MIRROR (#28)

• Proposed unified code: DESIRE_MIRROR
• Rationale: Identical function. DESIRE_MIRROR emphasizes the reflective technique (mirroring back what the audience already wants).

MERGE 18: C47 COMMON_MISTAKE + COMMON_ERROR (#39)

• Proposed unified code: COMMON_ERROR
• Rationale: Identical function. Minor naming difference (mistake vs. error).

MERGE 19: C48 DIFFERENTIATION + SOLUTION_DIFFERENTIATE (#40)

• Proposed unified code: SOLUTION_DIFFERENTIATE
• Rationale: Identical function. SOLUTION_ prefix clarifies what is being differentiated.

MERGE 20: C62 METHOD_SAFETY + METHOD_NATURAL (#31)

• Proposed unified code: METHOD_NATURAL
• Rationale: Both emphasize the method is natural and safe. The BR cluster includes safety language that goes slightly beyond "natural" (e.g., "sem efeitos colaterais"), but the core function is the same: reassuring the viewer about safety/naturalness.

MERGE 21: C22 PAIN_DEEPENING + PAIN_ARTICULATE (#5) + PAIN_AGITATE (#6)

• Proposed unified code: PAIN_DEEPEN (with optional sub-codes PAIN_ARTICULATE and PAIN_AGITATE)
• Rationale: C22 combines naming pain and amplifying it. If the unified taxonomy wants finer granularity, retain both reference sub-codes under a parent PAIN_DEEPEN code. If simplicity is preferred, merge all three into PAIN_DEEPEN.

MERGE 22: C24 AUDIENCE_QUALIFICATION + SELF_SELECT (#2) + QUALIFY (#33)

• Proposed unified code: AUDIENCE_QUALIFY (with optional sub-codes SELF_SELECT for early-funnel and QUALIFY for late-funnel)
• Rationale: C24 covers both the early "keep watching" qualifier and the late "who should buy" restatement. The reference vocabulary splits by position; the BR corpus merges by function. A unified code with positional sub-codes serves both needs.

SECTION 5: COVERAGE MATRIX

Summary Statistics

Verification: 24 MATCH + 18 PARTIAL + 3 SPLIT + 13 GAP_IN_REFERENCE = 58 (note: some clusters map to multiple statuses; the primary status for each of the 63 clusters sums correctly as follows)

Detail breakdown by cluster status assignment:

• MATCH (24): C01, C02, C03, C04, C05, C06, C08, C09, C11, C16, C18, C19, C20, C21, C23, C25, C26, C31, C33, C35, C36, C39, C40, C47
  • Wait -- recount: C01, C02, C03, C04, C05, C06, C08, C09, C11, C16, C18, C19, C20, C21, C23, C25, C26, C31, C33, C35, C36, C38, C39, C40, C45, C47, C48, C62 = 28

Let me recount precisely from the mapping table:

MATCH (28 clusters): C01, C02, C03, C04, C05, C06, C08, C09, C11, C16, C18, C19, C20, C21, C23, C25, C26, C31, C33, C35, C36, C38, C39, C40, C45, C47, C48, C62

PARTIAL (17 clusters): C07, C12, C14, C15, C17, C27, C28, C29, C32, C37, C43, C50, C53, C54, C55, C56, C58

SPLIT (3 clusters): C10, C22, C24

GAP_IN_REFERENCE (13 clusters): C13, C30, C34, C41, C42, C44, C46, C49, C51, C52, C57, C60, C61

Verification: 28 + 17 + 3 + 13 = 61. Remaining 2 clusters: C59 (PARTIAL) and C63 (PARTIAL) = 61 + 2 = 63. Correct.

Corrected Counts

Coverage Calculations

Reference codes with at least a MATCH or PARTIAL from BR clusters:

The following reference codes are covered (at least one BR cluster maps to them, whether MATCH, PARTIAL, or SPLIT):

Reference codes NOT covered (GAP_IN_BR):

1. SCROLL_STOP (#1)
2. CONSPIRACY_LEAN (#8)
3. IDENTITY_SPEAK (#9)
4. CREDIBILITY_SEED (#10)
5. TRANSITION_PIVOT (#22)
6. Emotional_Mobilization (#42)
7. Viral_Trigger (#44)
8. Pause-for-emphasis (#46)
9. Curiosity_Amplifier (#47)

Note: SELF_SELECT (#2), PROOF_SPECIFICITY (#24), PROMISE_TIMELINE (#26), CTA_EMOTIONAL (#35), and Identity_Close (#45) are partially/indirectly covered but were listed in Section 3 because no BR cluster maps to them as a primary function. For coverage calculation purposes, they receive partial credit.

Final Coverage Percentages

Coverage Summary

The 47-code reference vocabulary and 63-cluster BR taxonomy overlap substantially but not completely. 80.9% of reference codes have at least partial coverage from BR clusters. The 9 uncovered reference codes fall into three categories:

1. Platform-specific mechanics (SCROLL_STOP, Pause-for-emphasis): audiovisual/platform behaviors not visible in text-based annotation.
2. Meta-level / emergent properties (Emotional_Mobilization, Viral_Trigger, Curiosity_Amplifier): aggregate effects that emerge from combinations of individual moves rather than being discrete beats.
3. Identity/tribal techniques (IDENTITY_SPEAK, CONSPIRACY_LEAN, CREDIBILITY_SEED): subtle identity and credibility devices that the BR annotators did not isolate, possibly because Brazilian health/beauty DR relies less on these or because they were subsumed under broader clusters.

The 13 GAP_IN_REFERENCE clusters reveal that Brazilian DR employs techniques the reference vocabulary does not cover, particularly:

• Proof modalities (DEMONSTRATIVE_PROOF)
• Tonal devices (GOSSIP_INTIMATE_TONE, STEALTH_SELLING)
• Structural/format devices (BULLETS_LIST, GAP_TO_VSL, ANALOGY)
• Narrative devices (STORY_EMOTIONAL, PRODUCT_BUILDING, SCIENTIFIC_DISCOVERY)
• Funnel-specific functions (AUDIENCE_EXPANSION, FUNNEL_BRIDGE)
• Positive price framing (PRICE_BENEFIT)
• Focusing frames (THE_ONE_THING)

End of Phase 4 Cross-Reference Analysis