# Audit: Natural sugars in fruit are healthy while added sugars are poison (in equivalent amounts).

- **Generated:** 2026-03-28
- **Reader summary:** [proof.md](proof.md)
- **Proof script:** [proof.py](proof.py)

---

## Claim Specification

| Field | Value |
|-------|-------|
| Subject | Natural fruit sugars vs. added sugars at equivalent doses |
| Property | Compound claim: SC1 (fruit sugar healthy) AND SC2 (added sugar = poison at equivalent dose) |
| Operator | >= |
| Threshold | 3 (per sub-claim) |
| Proof direction | Partial (SC1 affirmed, SC2 disproved) |
| Operator note | SC1 evaluated by ≥3 sources confirming whole-fruit sugar consumption is healthy. SC2 evaluated by ≥3 sources confirming added sugar at equivalent-to-fruit doses is "poison" or acutely toxic — if instead ≥3 sources contradict this, SC2 is disproved. "Poison" interpreted literally as toxic or acutely harmful, not merely "bad in excess." |

*Source: proof.py JSON summary*

---

## Fact Registry

| ID | Key | Label |
|----|-----|-------|
| B1 | harvard_health | Harvard Health: natural and added sugars metabolized the same way; fruit healthy due to food matrix |
| B2 | the_conversation | The Conversation: same caloric content per gram regardless of source; fiber explains health difference |
| B3 | aha_added_sugars | American Heart Association: added sugars are empty calories; recommends limiting, not eliminating |
| B4 | pmc_are_all_sugars_equal | European Journal of Nutrition review: food matrix (fiber, polyphenols) drives physiological differences |
| B5 | pmc_rct_equivalent | RCT (PMC8277919): no meaningful cardiometabolic differences between equivalent added-sugar drinks and fruit sugar |
| A1 | — | SC1 verified source count (≥3 sources confirm fruit sugar healthy) |
| A2 | — | SC2 disproof source count (≥3 sources contradict 'poison at equivalent dose') |

*Source: proof.py JSON summary*

---

## Full Evidence Table

### Type A (Computed) Facts

| ID | Fact | Method | Result |
|----|------|--------|--------|
| A1 | SC1 verified source count | count(verified SC1 citations) = 4 | 4 of 4 confirmed |
| A2 | SC2 disproof source count | count(verified SC2-disproof citations) = 4 | 4 of 4 confirmed |

*Source: proof.py JSON summary*

### Type B (Empirical) Facts

| ID | Fact | Source | URL | Quote | Status | Method | Credibility |
|----|------|--------|-----|-------|--------|--------|-------------|
| B1 | Identical metabolism; fruit healthy via food matrix | Harvard Health Publishing, Harvard Medical School | https://www.health.harvard.edu/blog/are-certain-types-of-sugars-healthier-than-others-2019052916699 | "Natural and added sugars are metabolized the same way in our bodies." | Verified | full_quote | Tier 4 (academic) |
| B2 | Same calories regardless of source; fiber protective | The Conversation (peer-reviewed academic commentary) | https://theconversation.com/if-sugar-is-so-bad-for-us-why-is-the-sugar-in-fruit-ok-89958 | "All types of sugars will give us the same amount of calories, whether they are from fruit or soft drink." | Verified | full_quote | Tier 3 (major_news) |
| B3 | Added sugars: empty calories, limit recommended, not eliminate | American Heart Association | https://www.heart.org/en/healthy-living/healthy-eating/eat-smart/sugar/added-sugars | "Added sugars contribute zero nutritional benefit but often many added calories that can lead to overweight or obesity." | Verified | full_quote | Tier 2 (unclassified) |
| B4 | Food matrix (fiber, polyphenols) explains physiological differences | European Journal of Nutrition (2024 systematic review, PMC11329689) | https://pmc.ncbi.nlm.nih.gov/articles/PMC11329689/ | "Initial evidence implicates physical structure, energy density, fibre, potassium and polyphenol content, as explanations for some of the observed responses." | Verified | full_quote | Tier 5 (government/NIH) |
| B5 | 4-week RCT: no cardiometabolic difference at equivalent sugar doses | PMC8277919 — 4-week RCT on equivalent added vs. fruit sugars | https://pmc.ncbi.nlm.nih.gov/articles/PMC8277919/ | "Despite being asked to consume additional sugar (up to 1,800 additional kJ/d), there were no changes in weight, blood pressure or other cardiometabolic risk factors, except by uric acid, in any of the intervention groups." | Verified | full_quote | Tier 5 (government/NIH) |

*Source: proof.py JSON summary*

---

## Citation Verification Details

**B1 — Harvard Health Publishing**
- Status: verified
- Method: full_quote
- Fetch mode: live
- Coverage: N/A (full quote match)

**B2 — The Conversation**
- Status: verified
- Method: full_quote
- Fetch mode: live
- Coverage: N/A (full quote match)

**B3 — American Heart Association**
- Status: verified
- Method: full_quote
- Fetch mode: live
- Coverage: N/A (full quote match)

**B4 — European Journal of Nutrition (PMC11329689)**
- Status: verified
- Method: full_quote
- Fetch mode: live
- Coverage: N/A (full quote match)
- Note: PMC pages inject inline reference markers ([1], superscripts) that can degrade fragment matching; the shorter targeted quote avoids this issue.

**B5 — PMC8277919 (RCT)**
- Status: verified
- Method: full_quote
- Fetch mode: live
- Coverage: N/A (full quote match)

*Source: proof.py JSON summary*

---

## Computation Traces

```
SC1: verified source count vs threshold: 4 >= 3 = True
SC2 disproof: verified source count vs threshold: 4 >= 3 = True
```

*Source: proof.py inline output (execution trace)*

---

## Independent Source Agreement (Rule 6)

### SC1: Multiple independent sources confirm whole fruit is healthy

| Source | Status |
|--------|--------|
| harvard_health | verified |
| the_conversation | verified |
| aha_added_sugars | verified |
| pmc_are_all_sugars_equal | verified |

- Sources consulted: 4
- Sources verified: 4
- Independence note: Sources span academic institutions (Harvard Medical School), peer-reviewed journals (European Journal of Nutrition via PMC), established health advocacy organizations (AHA), and science journalism (The Conversation). All reach the same conclusion via independent reasoning and evidence bases.

### SC2 disproof: Multiple independent sources contradict 'poison at equivalent doses'

| Source | Status |
|--------|--------|
| harvard_health | verified |
| the_conversation | verified |
| aha_added_sugars | verified |
| pmc_rct_equivalent | verified |

- Sources consulted: 4
- Sources verified: 4
- Independence note: Sources include a clinical RCT (PMC8277919 — direct experimental evidence), a major health authority (AHA), and two independent academic/journalistic sources. The RCT provides the strongest SC2 disproof: it directly tested equivalent doses and found no meaningful cardiometabolic differences.

*Source: proof.py JSON summary*

---

## Adversarial Checks (Rule 5)

**Check 1: Is there a scientific mechanism distinguishing 'natural' fructose from 'added' fructose at the molecular level?**

- Verification performed: Searched PubMed and general web for "natural vs added fructose metabolic difference molecular mechanism 2024". Found no peer-reviewed evidence that the fructose molecule from fruit is chemically or metabolically different from isolated fructose. Harvard Health (B1) and The Conversation (B2) both confirm identical metabolism. All differences are food-matrix effects.
- Finding: No such mechanism exists. Fructose is fructose regardless of source. Supports SC1 disproof of SC2's premise.
- Breaks proof: No

**Check 2: Does any authoritative health body describe added sugars as 'poison' at equivalent-to-fruit doses?**

- Verification performed: Searched official sites of WHO (who.int), FDA (fda.gov), AHA (heart.org), NIH (nih.gov) for "poison" in the context of added sugar. Also searched "added sugar toxicity equivalent fruit dose." Found all authorities recommend LIMITING added sugars (AHA: <6% calories/day; WHO: <10% total energy; FDA: <10% daily value) but none describe them as poison.
- Finding: No health authority uses "poison" language for added sugars. The AHA's strongest characterization is "zero nutritional benefit." This directly disproves SC2.
- Breaks proof: No

**Check 3: Is there a controlled trial showing equivalent amounts of added sugar cause significantly more harm?**

- Verification performed: Searched PubMed for "RCT added sugar vs fruit sugar equivalent cardiometabolic." Found PMC8277919 (4-week RCT). Result: no significant differences in weight, blood pressure, or cardiometabolic risk factors between equivalent added-sugar (soft drink) and fruit-sugar groups — except uric acid in overweight males (gout risk marker, not acute toxicity).
- Finding: The best available RCT on equivalent doses contradicts SC2. Elevated uric acid in one subgroup does not constitute "poison" at fruit-equivalent doses.
- Breaks proof: No

**Check 4: Could removing fiber from fruit make its sugar behave like added sugar?**

- Verification performed: Searched "fruit juice vs whole fruit sugar health comparison fiber." Found consistent evidence that fruit juice has worse health profiles than whole fruit, supporting the food-matrix explanation for SC1.
- Finding: The health benefit of fruit sugar is due to food matrix, not the sugar itself. This undercuts the claim's implied premise that "natural sugar" has unique properties — but reinforces the SC2 disproof (if fiber removal makes fruit sugar behave like added sugar, the molecules are equivalent).
- Breaks proof: No

*Source: proof.py JSON summary*

---

## Source Credibility Assessment

| Fact ID | Domain | Type | Tier | Note |
|---------|--------|------|------|------|
| B1 | harvard.edu | academic | 4 | Academic domain (.edu) |
| B2 | theconversation.com | major_news | 3 | Major news organization (academic editorial model) |
| B3 | heart.org | unclassified | 2 | Unclassified domain — verify source authority manually |
| B4 | nih.gov | government | 5 | Government domain (.gov) — NIH PubMed Central |
| B5 | nih.gov | government | 5 | Government domain (.gov) — NIH PubMed Central |

Note: B3 (heart.org, tier 2) is the only unclassified-domain source. The American Heart Association is a major, well-established health nonprofit, and this credibility tier reflects an automated domain classifier limitation rather than source unreliability. B3's claim is independently corroborated by B1 (Harvard), B4 (peer-reviewed journal), and B5 (clinical RCT) — all from higher-tier sources. The proof conclusions do not depend solely on B3.

*Source: proof.py JSON summary*

---

## Extraction Records

This is a qualitative consensus proof. No numeric values were extracted. Extraction records document citation verification status per source.

| ID | Quote Snippet | Verified | Countable |
|----|---------------|----------|-----------|
| B1 | "Natural and added sugars are metabolized the same way in our bodies." | verified | Yes |
| B2 | "All types of sugars will give us the same amount of calories, whether they are f..." | verified | Yes |
| B3 | "Added sugars contribute zero nutritional benefit but often many added calories t..." | verified | Yes |
| B4 | "Initial evidence implicates physical structure, energy density, fibre, potassium..." | verified | Yes |
| B5 | "Despite being asked to consume additional sugar (up to 1,800 additional kJ/d), t..." | verified | Yes |

All 5 citations are fully verified (full_quote method, live fetch). No partial or failed citations.

*Source: proof.py JSON summary*

---

## Hardening Checklist

| Rule | Status | Notes |
|------|--------|-------|
| Rule 1: No hand-typed extracted values | PASS | Qualitative proof — no numeric values extracted from quotes |
| Rule 2: Citations verified by fetching | PASS | All 5 URLs fetched live; all quotes confirmed (full_quote method) |
| Rule 3: Anchored to system time | PASS | date.today() used; PROOF_GENERATION_DATE = 2026-03-28 |
| Rule 4: Explicit claim interpretation | PASS | CLAIM_FORMAL with operator_note; "poison" interpreted literally; "equivalent amounts" explicitly scoped |
| Rule 5: Adversarial checks independent | PASS | 4 adversarial checks; each searched for counter-evidence, not restatements of the proof |
| Rule 6: Cross-checks independently sourced | PASS | SC1 uses 4 independent sources across 3 institution types; SC2 disproof includes an RCT independent of SC1 sources |
| Rule 7: No hard-coded constants or inline formulas | PASS | compare() used for all claim evaluations; no inline math |
| validate_proof.py | PASS (1 warning) | 14/15 checks passed; warning: sc2_holds uses logical inversion of compare() result — correct but not directly a compare() call |

*Source: proof.py inline output (execution trace) and validate_proof.py results*

---

*Generated by [proof-engine](https://github.com/yaniv-golan/proof-engine) v0.10.0 on 2026-03-28.*