VALIDATION SECTION

Taxonomy v1.0 — Merge/Removal Decisions & Discriminant Cross-Check

1. CODES MERGED (and why)

2. CODES REMOVED (and why)

3. DISCRIMINANT DIFFERENTIATION: CONFUSABLE PAIRS

SPOILER_MUP vs. MUP_EXPLAIN

• SPOILER_MUP: "Is this sentence MENTIONING or HINTING at a hidden cause... WITHOUT fully explaining how it works?"
• MUP_EXPLAIN: "Is this sentence providing a COMPLETE explanation of HOW the problem mechanism works — with a cause-and-effect chain?"
• Test sentence: "Scientists found a toxic protein that eats your cartilage." → SPOILER_MUP ✅ (hints at cause, no chain). MUP_EXPLAIN ❌ (no cause-and-effect chain).
• Test sentence: "This protein attaches to the cartilage surface, breaks down collagen fibers, triggers inflammation, which accelerates bone erosion." → MUP_EXPLAIN ✅ (full chain). SPOILER_MUP ❌ (fully explained).

SPOILER_MSOL vs. SOLUTION_PRESENT vs. MSOL_EXPLAIN

• SPOILER_MSOL: "HINTING at or BRIEFLY MENTIONING the solution WITHOUT explaining how or why it works?"
• SOLUTION_PRESENT: "FORMALLY NAMING or INTRODUCING the solution for the first time — giving it a specific identity?"
• MSOL_EXPLAIN: "Providing a COMPLETE mechanistic explanation of HOW the solution works?"
• Test: "There's a simple morning trick using a common spice..." → SPOILER_MSOL ✅ (hint, no name).
• Test: "It's called NeuroCalm — developed at Johns Hopkins." → SOLUTION_PRESENT ✅ (formal naming). SPOILER_MSOL ❌ (has identity). MSOL_EXPLAIN ❌ (no mechanism).
• Test: "NeuroCalm works by binding to the receptor and blocking the inflammatory signal cascade." → MSOL_EXPLAIN ✅ (full mechanism).

ENEMY_FRAME vs. INVALIDATE_SOLUTIONS

• ENEMY_FRAME: "Naming an EXTERNAL ENTITY that is deliberately harming, deceiving, or profiting?"
• INVALIDATE_SOLUTIONS: "Stating a SPECIFIC method, product, or approach does NOT work?"
• Test: "Big Pharma doesn't want you to know about this." → ENEMY_FRAME ✅ (WHO is blamed). INVALIDATE_SOLUTIONS ❌ (no specific product attacked).
• Test: "Advil and Tylenol only mask the symptoms." → INVALIDATE_SOLUTIONS ✅ (WHAT doesn't work). ENEMY_FRAME ❌ (no villain entity blamed).
• Edge case: "The pharmaceutical industry keeps selling you drugs they know don't work." → BOTH could apply. Primary: ENEMY_FRAME (entity blamed). Secondary: INVALIDATE_SOLUTIONS (drugs don't work). In dual-tag scenarios, assign both.

FEAR_DEEPEN vs. PAIN_ARTICULATE vs. NEGATIVE_FUTURE_PACE

• PAIN_ARTICULATE: "NAMING a specific pain the viewer currently lives with?"
• FEAR_DEEPEN: "ESCALATING fear by describing a consequence WORSE than what the viewer was already aware of?"
• NEGATIVE_FUTURE_PACE: "Describing what will happen IF THEY DO NOT take action?"
• Test: "You wake up every morning with stiff, aching joints." → PAIN_ARTICULATE ✅ (names current pain). FEAR_DEEPEN ❌ (no escalation). NEGATIVE_FUTURE_PACE ❌ (no future projection).
• Test: "What most people don't realize is that this inflammation can also trigger heart attacks." → FEAR_DEEPEN ✅ (escalates to worse consequence). PAIN_ARTICULATE ❌ (not naming current experience). NEGATIVE_FUTURE_PACE ❌ (not conditional on inaction).
• Test: "If you don't address this now, in 5 years you could lose the ability to walk." → NEGATIVE_FUTURE_PACE ✅ (conditional future if no action). FEAR_DEEPEN ❌ (no new information, just projection). PAIN_ARTICULATE ❌ (not current pain).

PROOF_SOCIAL vs. PROOF_PERSONAL vs. PROOF_SCIENTIFIC

• PROOF_SOCIAL: "Citing experiences of OTHER PEOPLE — through numbers, testimonials, or named individuals?"
• PROOF_PERSONAL: "The SPEAKER sharing their OWN personal experience — using first-person 'I'?"
• PROOF_SCIENTIFIC: "Citing a STUDY, CLINICAL TRIAL, RESEARCH INSTITUTION, or SPECIFIC STATISTICAL DATA?"
• Test: "Over 127,000 people have used this." → PROOF_SOCIAL ✅. PROOF_PERSONAL ❌. PROOF_SCIENTIFIC ❌.
• Test: "I personally lost 40 pounds using this method." → PROOF_PERSONAL ✅. PROOF_SOCIAL ❌ (speaker, not others). PROOF_SCIENTIFIC ❌.
• Test: "A 2023 Stanford study found 87% reduction." → PROOF_SCIENTIFIC ✅. PROOF_SOCIAL ❌ (study, not people stories). PROOF_PERSONAL ❌.

CTA_MECHANICAL vs. CTA_BUILD

• CTA_BUILD: "PREPARING the viewer for the CTA — describing what they'll see or framing the moment — WITHOUT the direct instruction?"
• CTA_MECHANICAL: "Giving a DIRECT INSTRUCTION to take a specific physical action?"
• Test: "On the next page, you'll see everything you need." → CTA_BUILD ✅ (describes what's next). CTA_MECHANICAL ❌ (no instruction).
• Test: "Click the button below right now." → CTA_MECHANICAL ✅ (direct instruction). CTA_BUILD ❌ (no framing, pure command).

SCARCITY_INFO vs. SCARCITY_OFFER

• SCARCITY_INFO: "Claiming the INFORMATION or VIDEO may be removed or censored?"
• SCARCITY_OFFER: "Claiming the OFFER, PRICE, or AVAILABILITY is limited?"
• Test: "This video may be taken down at any moment." → SCARCITY_INFO ✅. SCARCITY_OFFER ❌.
• Test: "We only have 500 bottles left." → SCARCITY_OFFER ✅. SCARCITY_INFO ❌.

CURIOSITY_OPEN vs. INVALIDATE_BELIEF

• CURIOSITY_OPEN: "Creates a specific UNANSWERED QUESTION or information gap?"
• INVALIDATE_BELIEF: "States something the viewer BELIEVES is actually WRONG?"
• Test: "What does a lunch lady have to do with wrinkles?" → CURIOSITY_OPEN ✅ (question creating gap). INVALIDATE_BELIEF ❌ (no belief attacked).
• Test: "Fiber is actually making your constipation worse." → INVALIDATE_BELIEF ✅ (breaks belief). CURIOSITY_OPEN ❌ (no open question, it's a statement).
• Edge case: "What if everything you've been taught about joint pain is completely wrong?" → BOTH. Primary: INVALIDATE_BELIEF (challenges belief). Secondary: CURIOSITY_OPEN (creates question). Assign both if applicable.

INVALIDATE_SOLUTIONS vs. SOLUTION_DIFFERENTIATE

• INVALIDATE_SOLUTIONS: "Stating a method DOES NOT WORK or is insufficient?"
• SOLUTION_DIFFERENTIATE: "CONTRASTING the product with alternatives to show SUPERIORITY on a feature?"
• Test: "Those creams can't penetrate deep enough." → INVALIDATE_SOLUTIONS ✅ (doesn't work). SOLUTION_DIFFERENTIATE ❌ (no contrast with own product).
• Test: "Unlike regular olive oil, ours retains 10x the polyphenols." → SOLUTION_DIFFERENTIATE ✅ (superiority claim). INVALIDATE_SOLUTIONS ❌ (doesn't say alternatives fail).

METHOD_SIMPLE vs. RAPID_RESULT

• METHOD_SIMPLE: "Emphasizing how EASY or EFFORTLESS the method is?"
• RAPID_RESULT: "Emphasizing HOW FAST results appear with a concrete timeframe?"
• Test: "All you do is mix one teaspoon into water — that's it." → METHOD_SIMPLE ✅ (ease). RAPID_RESULT ❌ (no time claim).
• Test: "You'll see results in just 7 days." → RAPID_RESULT ✅ (speed with timeframe). METHOD_SIMPLE ❌ (no ease claim).
• Edge case: "In 30 seconds a day, you'll see results within a week." → BOTH. METHOD_SIMPLE (30 seconds = easy). RAPID_RESULT (within a week = fast).

PROMISE vs. FUTURE_PACE vs. BENEFITS

• PROMISE: "Stating a SPECIFIC end result or transformation the viewer WILL achieve?"
• FUTURE_PACE: "Asking the viewer to IMAGINE their positive future state?"
• BENEFITS: "Listing specific positive outcomes as enumerated selling points?"
• Test: "You will be pain-free within 30 days." → PROMISE ✅ (core transformation). FUTURE_PACE ❌ (not asking to imagine). BENEFITS ❌ (core promise, not enumerated list).
• Test: "Imagine waking up full of energy, pain-free." → FUTURE_PACE ✅ (visualization). PROMISE ❌ (imagining, not promising). BENEFITS ❌ (not listing).
• Test: "It boosts metabolism, reduces inflammation, and improves sleep." → BENEFITS ✅ (enumerated list). PROMISE ❌ (not the core commitment). FUTURE_PACE ❌ (not visualization).

WORKS_FOR_ALL vs. AUDIENCE_EXPAND

• WORKS_FOR_ALL: "Claiming the method works for EVERYONE using universal language?"
• AUDIENCE_EXPAND: "Listing ADDITIONAL problems or conditions BEYOND the primary one?"
• Test: "This works for anyone, regardless of age or body type." → WORKS_FOR_ALL ✅ (universal claim). AUDIENCE_EXPAND ❌ (no new conditions listed).
• Test: "It's not just for joint pain — it also helps with back pain, stiffness, and arthritis." → AUDIENCE_EXPAND ✅ (new conditions listed). WORKS_FOR_ALL ❌ (not universal claim, specific conditions).

4. UNCERTAINTY FLAGS

STORY_TRANSPORT — position ambiguity

The code appears in early position but can extend across many sentences. As a delivery VEHICLE rather than a discrete beat, its annotation may overlap with PAIN_ARTICULATE (story about pain) or PROOF_PERSONAL (personal story as proof). Resolution: STORY_TRANSPORT should be tagged when the PRIMARY function of the sentence is NARRATIVE TRANSPORT (telling a story) rather than pain articulation or proof delivery. If the story is clearly serving as proof of the product working, tag PROOF_PERSONAL instead. If it's articulating pain through narrative, tag PAIN_ARTICULATE.

VALUE_FRAME vs. INVALIDATE_ON_PRICE — boundary cases

VALUE_FRAME positively frames OWN price. INVALIDATE_ON_PRICE attacks competitor price. When a sentence does BOTH ("Unlike $500 surgeries, this is just $49"), it should receive both tags. The discriminants are designed to be complementary, not mutually exclusive, in this case.

SELF_SELECT — absorbed too many codes?

SELF_SELECT absorbed QUALIFY, IDENTITY_SPEAK, and IDENTITY_CLOSE. The concern is that qualification early in the ad serves a different strategic purpose than qualification near the CTA. Resolution: The typical_position field is "first" because that's the FIRST typical appearance, but the code can appear at any position. Annotators should note position context. A future v2.0 could re-split if position data reveals distinct functions.

HOPE_BRIDGE — absorbed SOLUTION_2_0_BRIDGE

The "1.0 → 2.0" pattern is culturally specific to Brazilian DR and represents a distinct rhetorical move (explicitly naming the upgrade). Absorbing it into HOPE_BRIDGE may lose this nuance. Resolution: For v1.0, the merge is justified because the structural function (pivot from problem to solution) is identical. If Brazilian-specific analysis needs to distinguish "generic hope bridge" from "explicit 2.0 upgrade," a future version could re-introduce it.

5. FINAL CODE COUNT BY CATEGORY

End of Validation Section Taxonomy v1.0 — 50 operational codes from 78 analytical codes 78 → 50: 21 merges, 7 removals