Human Alimentation: A Multidimensional Feeding Meta-Synthesis

Synthesis model: GPT 5.6 Sol

Opening: Executive Synthesis and Central Thesis

Human alimentation is not a contest to identify one biologically correct menu. It is the continuing process by which people and communities obtain, transform, share, tolerate, metabolize, interpret, and govern food under changing ecological, historical, physiological, cultural, and material conditions.^6:supp^7:supp^8:rel^9:supp^10:rel Archaeological methods recover different slices of past eating rather than a complete ancestral menu (method inventory); contemporary foodways encode relationships with land, seasons, migration, identity, adaptation, and resilience (UNESCO account); and modern dietary-pattern scores remain investigator-made abstractions that can omit foods and fail to capture actual intake complexity (classification limitation). These independent scopes converge on a central thesis: there is no evidence-supported universal human diet in this corpus, but there are multiple potentially valid dietary paths whose suitability depends on explicit goals, substitutions, food matrices, life circumstances, observed response, and evidence quality.^11:rel^12:supp^13:supp^14:supp^15:supp

This meta-synthesis uses realized nourishment as its organizing construct. This is this meta-synthesis's own cross-scope hypothesis, not a term asserted by any one source: nourishment is realized only when food is sufficiently available and affordable (price pressure), physically and administratively accessible (disability-access account), preparable with available time and infrastructure (vendor-capacity questions), nutritionally adequate and bioavailable (fermentation mechanism), physiologically tolerated (IBD participation finding), culturally and religiously usable (religious preference guidance), socially sustainable, and meaningfully chosen under legitimate individual and community control (food-sovereignty definition).^16:supp^17:supp^18:supp^19:supp^4:rel This hypothesis should be tested through research that measures those domains jointly rather than treating access, culture, physiology, and governance as exceptions to nutrient optimization.

The evidence supports several strong bounded conclusions.^20:rel^21:rel^22:rel^23:rel^24:rel Protein sources materially differ in amino-acid profile, digestibility, co-nutrients, preparation requirements, cost structure, cultural meaning, and production system; high digestibility is useful for adequacy questions but does not itself establish long-term health superiority (comparative scores; outcome boundary). Food structure and transformation alter what is released and absorbed: whole and ground nuts yield different absorbed energy (matrix observation), while specific fermentations can increase iron or zinc availability (iron; zinc).^25:supp^26:rel^27:supp^28:supp^29:supp Diet-outcome evidence is substitution-specific: healthful and unhealthful plant indices move in opposite directions for diabetes (quality reversal), and low-carbohydrate mortality associations differ by plant versus animal replacement sources (plant replacement; animal replacement). Many named diets improve short-term markers, but advantages commonly diminish by twelve months (durability result).^30:supp^31:supp^32:supp^33:supp^34:supp

The corpus also strongly supports heterogeneity. Low-FODMAP sources report substantial digestive-symptom benefit while preserving a meaningful nonresponder group (response estimate); first-person records include benefit, no identifiable triggers, failed elimination, psychological harm, and social burden (benefit; no identifiable triggers; failed elimination; tracking harm). Such reports are attributed individual evidence and hypothesis generators.^20:rel^35:supp^36:supp^22:rel^21:rel They are neither population-level causal proof nor disposable noise.

The practical result is a multidimensional framework, not a ranked diet list.^37:rel^13:supp^38:rel^39:rel It asks: What outcome matters? What food or practice is actually being replaced? Can adequacy be maintained? What does preparation change? Which constraints bind? What evidence type supports the claim? What response is observed, over what time?^40:supp What harms, burdens, and opportunity costs arise? Which interests shaped the evidence and recommendation? When should the path be continued, adapted, broadened, or stopped?

Scope and Boundaries

This document integrates nine investigations: contemporary and global foodways; ancient and historical diets; protein sources; nutrients, matrices, processing, hydration, and preparation; disease, mortality, lifespan, and healthy aging; mood, cognition, sleep, wellbeing, digestion, and reported inflammation; life stage, sex-related physiology, genetics, microbiome, disability, and access; lived and community evidence; and methods, governance, conflicts, and incentives.

It does not provide personalized medical advice, diagnose deficiency or disease, prescribe therapeutic restriction, or identify a lifespan-maximizing diet.^41:rel It does not treat official guidance as self-validating, academic publication as inherently superior, community knowledge as automatically causal, or outsider status as proof of independence. It preserves evidence-type boundaries:

Material and archaeological evidence can establish recovered residues, tissues, objects, and documented practices, while dietary proportions and meaning require reconstruction. Bone isotopes can indicate major long-term protein sources, whereas dental calculus can retain particular particles from food, medicine, hygiene, or environment (calculus evidence).
Composition and digestion studies can establish nutrient content, release, or acute availability, but not durable adherence, disease prevention, or cultural feasibility.
Observational studies can estimate associations under measurement and adjustment assumptions, but cannot guarantee removal of residual confounding.
Interventions can strengthen causal inference for the assigned exposure and observed period, while adherence, attrition, blinding, comparator quality, and short duration still limit interpretation (trial-validity markers).
Mechanisms and biomarkers can test plausibility but do not automatically establish clinical events, function, quality of life, or survival (endpoint distinction).
Guidelines and regulations establish institutional decisions, thresholds, and procedures, not universal biological optima or comparative benefit.
Cultural and community knowledge directly establishes meanings, practices, priorities, continuity, and locally identified harms; biomedical translation requires a separately appropriate design.
Journalism and practitioner accounts can preserve ecological detail, implementation knowledge, and hypotheses while remaining vulnerable to selection, service incentives, and narrative compression.
Industry and advocacy claims can contain testable data and domain expertise while carrying commercial, political, or ideological incentives that require symmetric scrutiny.
First-person observations directly establish what a person reports experiencing and attributing. Causal generalization requires prospective controls, comparison, replication, or triangulation appropriate to the claim.^11:rel^20:rel^42:rel^21:rel^43:rel

The evidence is geographically and demographically uneven. It includes Indigenous communities, Kenyan heritage work, Mediterranean research, CHamoru adaptation, North American refugee and disability access, archaeological sites in Europe, Asia, Africa, and southwest Asia, and predominantly US-centered nutrition studies and governance.^44:rel^45:rel^46:supp^47:rel^48:supp It is not a representative census of world diets. Evidence for ancestry-based individual prediction is particularly thin; evidence for children, pregnancy, older adults, disability, and many clinical populations is uneven; and emerging proteins have limited long-term product-specific follow-up.^49:rel^50:rel^51:rel^52:rel

Sub-Domain I: Diets Across Ecology, Culture, History, and Access

No single ancestral baseline

Ancient evidence undermines both a universal meat-only past and a universal plant-rich past.^53:supp^54:rel^55:rel^56:rel Isotopic summaries characterize some European Neanderthals as high-trophic consumers and can indicate dominant lifetime protein sources (high-trophic interpretation), while calculus microfossils, charred plant mixtures, and residues document plant contact, cooking, and multi-step preparation (microfossils; charred mixture; preparation interpretation). These findings can coexist because isotopes and residues represent different exposure windows and dietary components.^57:rel^58:supp^59:rel^60:rel They support human dietary flexibility, not one reconstructed prescription.

The farming transition is similarly plural.^61:rel^62:rel^63:rel^64:rel^65:rel Calculus evidence has been interpreted as domesticated-plant consumption among Mesolithic Balkan foragers (Balkan evidence), while African microbotanical evidence documents domestic cereals around 7,000 years ago (African cereal evidence). Scholarship has also advanced a declining-health thesis for early agriculture (declining-health thesis).^66:rel^67:supp The strongest integrated reading is not that farming was harmless or uniformly harmful, but that ingredient adoption could be gradual while labor, infection, inequality, land control, settlement, and health changed sharply in some settings.

Historical recipes and archives reveal technical complexity without establishing population prevalence.^56:rel^68:rel^69:rel Medieval texts document multi-stage mushroom and pickling procedures (mushroom recipe; pickle procedure), while modern museum recipes explicitly adapt originals for contemporary kitchens (adaptation statement). Archives preserve cookbooks, diaries, government files, songs, and photographs (holdings overview, but survival reflects collectors, literacy, bureaucracy, property, and institutional acquisition.^70:supp^71:rel^72:supp^73:rel The past can expand possibilities and generate hypotheses; it cannot select a present menu without present outcome, safety, access, and cultural evidence.

Foodways as adaptation and power

Contemporary foodways are lived systems, not nutrient delivery channels.^8:rel^9:supp^74:supp^44:rel^75:rel Community sources describe food as preserving knowledge, memory, and tradition (community account, while CHamoru testimony describes family cooking and language sustaining continuity even after colonial ingredients such as Spam, rice, and eggs entered daily practice (adaptation testimony). These accounts support two simultaneous propositions: protecting land, species, knowledge, and practices can be essential; and continuity need not require ingredient purity.^75:rel^44:rel^76:supp^77:supp^8:rel

Structural disruption is directly documented more strongly than generic suspicion. Historical institutional accounts attribute disruption of Indigenous food systems to land theft and forced removal (historical account); Indigenous testimony adds rights loss, environmental toxins, cultural oppression, and practical barriers to traditional foods (educator testimony). Kenyan heritage work reports modernization and urbanization pressure toward abandonment of traditional foodways (transition claim).^78:supp^79:supp^8:rel^76:supp^80:supp These sources differ in motive and method but converge on mechanisms of disruption.

Food security and food sovereignty answer different questions. Food insecurity concerns insufficient affordable nutritious food because of resource constraints (definition); food sovereignty centers who controls land, production, distribution, policy, and knowledge (movement definition).^81:supp^17:supp^82:rel^83:supp^84:supp A system can supply calories while weakening control, or expand local control without yet securing sufficient food. Indigenous organizations explicitly join culturally appropriate food to hunting, fishing, gathering, growing, provider compensation, distribution, and treaty rights (community framework). Neither sufficiency nor authority should be treated as a proxy for the other.

Access changes the intervention

Material access is not a compliance issue added after identifying an optimal diet.^85:supp^16:supp It determines which diet exists. Disability-focused evidence identifies transport, physical store access, affordability, and availability of condition-compatible foods as part of the food environment (disability-access account). Food insufficiency was reported at 24% among disabled people versus 9% among non-disabled people in one June 2023 analysis (disability disparity).^86:supp^87:supp^88:rel^89:rel^90:rel Refugee-serving organizations identify food deserts, inflation, transit, and the price of culturally appropriate produce as barriers (newcomer barriers; structural account).

Clinical guidance makes the causal chain concrete: discussing nutrition with a pregnant person losing weight after job loss is described as meaningless without food-assistance referral (clinical example, and housing support may enable gestational-diabetes management more effectively than repeated monitoring advice when equipment is lost during moves (housing example).^91:supp^92:supp^93:supp^94:supp^95:rel The graph independently links food insecurity to housing, healthcare, employment, clinical practitioners, hunger, and malnutrition. Agent hypothesis, not asserted by any source: this bridge suggests that the marginal value of biological precision can approach zero when the binding constraint is provision, housing, equipment, or access. A discriminating trial would compare biological personalization alone, structural support alone, and their combination.

Convenience has nutritional and labor effects. Public-health accounts characterize ultra-processed foods as shelf-stable, affordable, and able to save cooking time, effort, and mental energy (convenience account).^96:supp^97:supp^98:supp^99:rel^100:rel The same literature identifies time, accessibility, and price as barriers to reducing them (barrier statement). Reform that ignores replacement cost, storage, fuel, disability, and caregiver labor can externalize work back onto households even when its nutritional objective is reasonable.^16:supp^101:rel^102:rel^103:supp^104:supp

Sub-Domain II: Nutritional Architecture

Protein: adequacy is a portfolio property^105:supp^106:supp^107:rel^108:supp

Protein decisions require four separate layers: composition, digestibility, total-diet adequacy, and outcomes. Nine amino acids are classified as indispensable because diet must supply them (amino-acid definition).^109:supp^110:rel^111:rel^112:rel^113:rel Reported DIAAS values vary widely: milk protein concentrate 141, beef 112, chicken 108, egg 101, soy flour 105, potato 100, pea concentrate 73, and insect protein 75 (comparative table; beef; soy; potato). This supports source- and processing-specific variation, not a strict animal-versus-plant hierarchy.^114:supp^115:supp

Animal foods can provide concentrated, highly digestible protein and co-nutrients in modest volumes; that practical margin may matter for low appetite, high requirements, recovery, or limited variety. Diverse plant foods can also supply substantial protein: repository examples report roughly 14-16 g per cup of beans, 18 g per cup of lentils, 22-23 g in tofu or edamame, and 27 g per 150 g of tempeh (serving examples).^116:supp^117:supp^118:rel^119:rel^108:supp The unresolved question is whether score differences become real-world inadequacy under actual energy intake, serving capacity, preparation, life stage, and tolerance.

Aquatic foods are not one category. A database spanning more than 3,750 species records minerals, vitamins, and fatty acids (composition database); aquatic animal foods can supply protein alongside zinc, iron, selenium, iodine, vitamins A, B6, B12 and D, and long-chain omega-3 fats (co-nutrient account).^120:supp^121:supp^122:rel^123:rel^124:rel Species-specific mercury and pathogen risks require context (mercury guidance; temperature control). Fish-centered animal intake may also differ materially from processed-meat-centered patterns; cohort authors explicitly caution against transferring their animal-source findings to cultures where fish is the principal meat (generalizability warning).^125:supp^126:rel^49:rel^50:rel

Fungal, insect, fermented, and emerging proteins enlarge the portfolio without proving drop-in equivalence. Sixteen controlled mycoprotein trials included 432 participants and suggested the strongest evidence for lower total cholesterol and short-term energy intake, with less conclusive glucose and insulin findings (review scope; stronger endpoints; uncertain endpoints).^127:supp^128:supp^129:supp^130:supp^131:supp A lesser-mealworm crossover trial demonstrated rapid digestion and amino-acid absorption in healthy young men (acute trial). Precision fermentation can manufacture target molecules similar in constitution to conventional counterparts (production claim), but molecular similarity does not establish identical matrix, allergenicity, affordability, or long-term outcome.^132:supp^133:rel^134:rel^135:rel^136:rel

Agent hypothesis, not asserted by any source: protein efficiency has multiple denominators: usable amino acids per serving, dollar, hectare, unit of energy, minute of household labor, and culturally acceptable meal. These rankings can conflict. The hypothesis follows from digestibility variation (scores), pulse affordability claims (beans), preparation-dependent tempeh preservation (tradeoff), and energy-dependent cultivated-meat impacts (energy dependence).^108:supp^137:supp

Fats: source and replacement are part of the exposure

Saturated fat can raise LDL under controlled conditions, and the American Heart Association interprets this as a cardiovascular pathway (feeding result; AHA interpretation).^138:supp^139:supp^140:supp^141:supp^142:contr A Cochrane synthesis reported 17% fewer cardiovascular events across thirteen studies when saturated fat was reduced and rated the evidence moderate (intervention synthesis). Yet a large observational synthesis did not find robust associations with mortality or several major endpoints and emphasized the comparison nutrient (endpoint synthesis).^143:supp^144:supp^145:supp

These findings need not be forced into a binary verdict. Replacement with polyunsaturated fat was associated with lower cardiovascular risk (substitution association); replacement with high-glycaemic carbohydrate was associated with higher risk, while lower-glycaemic fruits, vegetables, pulses, and grains were associated with lower risk (quality-specific replacement). Dairy-matrix research proposes that saturated-fat content alone may not predict food-level effects (matrix hypothesis), but that hypothesis does not establish that every dairy food is protective.^146:supp^147:supp^148:supp^149:rel^150:rel A useful decision specifies the source food, replacement, baseline risk, endpoint, and observed response.

Carbohydrates and fibre: quantity is not quality^151:supp

Carbohydrate is metabolized into glucose, but total carbohydrate share cannot identify food quality, fibre, processing, or replacement (basic metabolism). Cohort evidence reports a U-shaped association between carbohydrate share and mortality, with both low and high intake associated with greater mortality than moderate intake (U-shape).^152:supp^153:supp^154:supp^155:supp^156:rel The more discriminating observation is that replacement source changed direction: plant-derived fat and protein replacement was associated with lower mortality, animal-derived replacement with higher mortality (plant; animal). This remains observational, but it shows why “low carbohydrate” is not one exposure.^157:supp^158:rel^159:supp^160:supp^161:rel

Fibre links glucose, bowel function, satiety, whole grains, micronutrients, microbiome, and processing. Randomized evidence summarized in the corpus reports improved post-meal and longer-term glucose metabolism with soluble or viscous fibre (intervention claim).^162:supp^163:rel Condition-specific guidance distinguishes soluble from insoluble fibre in IBS rather than treating more fibre as universally better (IBS guidance). Type, dose, fluid, baseline diet, gastrointestinal condition, and adaptation all modify practical effect.^164:rel^165:rel^166:supp^162:supp^167:supp

Micronutrients, hydration, and food matrix

Micronutrient composition is not equivalent to availability.^168:supp^28:supp^25:supp^29:supp Simulated digestion methods estimate release from matrices (INFOGEST method), while fermentation examples show food-specific increases in iron and zinc accessibility (vegetables; cassava). Restriction can create new inadequacy: calorie restriction in overweight adults improved selected metabolic measures while nutrient adequacy fell during the restriction phase before recovering during stabilization (adequacy tradeoff); a vegetarian author reported iron deficiency requiring monitoring and supplementation (individual harm signal).^169:supp^170:supp^171:supp^172:rel^173:rel

Hydration is essential but thinly evidenced in this corpus. CDC guidance associates dehydration with unclear thinking, mood change, overheating, constipation, and kidney stones (public-health claim).^174:supp^175:supp^176:rel^177:rel It does not establish one intake target across climate, exertion, illness, food water, medication, or renal function. Hydration should be monitored as a contextual variable, especially when fibre, heat, activity, vomiting, diarrhea, or medications change.^178:rel^179:supp^180:rel

The food matrix makes structure part of the exposure. Whole nuts reportedly yield about 30% less absorbed energy than ground nuts because fibre and fat remain less accessible, while some fat-soluble-vitamin absorption also falls (energy release; tradeoff).^181:supp^182:supp^183:rel^184:rel^25:supp Milk can become liquid, gelled yogurt, or solid cheese through heating, fermentation, and ripening (forms; transformation). Matrix awareness does not mean intact is always better; cooking and fermentation can improve safety and accessibility while reducing or altering other nutrients.^185:supp^25:supp^26:rel^184:rel^186:rel

Preparation, fermentation, and processing

Preparation links safety, preservation, palatability, bioavailability, labor, and identity.^187:rel^188:rel Cooking can improve digestion and availability of some nutrients while reducing others (availability; loss). Nutrients leached into boiling water may remain in the meal when broth is consumed (cooking-liquid context).^189:supp^190:supp^191:rel^192:rel^193:rel No method preserves or maximizes every nutrient.

Fermentation historically extended seasonal storage and can inhibit spoilage or pathogens through controlled acidification (storage function; safety mechanism).^194:supp^195:supp^196:supp^197:supp^198:supp It also carries taste, ritual, heritage, and memory (cultural observation). Fermented is not synonymous with probiotic: products may contain no live organisms at consumption or insufficient amounts for a demonstrated host benefit (qualification).^199:supp^200:supp^201:supp^202:supp^203:supp Substrate, organisms, process, storage, salt, sugar, batch, and host must be specified.

Ultra-processing is likewise neither a nutrient nor one homogeneous exposure.^204:rel^205:supp^99:rel^206:supp^207:rel A controlled 20-person crossover feeding trial found roughly 500 kcal/day greater intake and about two pounds of weight gain during the tested ultra-processed package over fourteen days per condition (trial result). That supports causality for the tested packages, not every NOVA Group 4 product.^208:rel^209:rel^210:rel^211:rel^212:supp Some classified whole-grain breads and yogurts are associated with lower chronic-disease risk (subtype exception), and a review argues that subtype heterogeneity blocks one class-wide causal effect (causal critique).

Agent hypothesis, not asserted by any source: “controlled transformation” is more discriminating than “natural versus processed.” The relevant questions are what a transformation releases, removes, adds, preserves, standardizes, accelerates, or makes feasible, and for whom.^213:rel^214:rel^215:rel^216:supp^100:rel This hypothesis is grounded in nut structure (matrix result), fermentation-mediated mineral access (mechanism), cooking tradeoffs (loss), and the ultra-processed feeding result (package effect).

Sub-Domain III: Outcomes Across Lifespan, Disease, Function, and Wellbeing^217:rel^218:rel^219:rel^220:rel^221:rel

Mortality and healthy aging

Direct lifespan evidence for a specific diet is thin. Long-horizon cohorts provide associations; interventions usually measure months of weight, symptoms, or biomarkers rather than mortality.^222:supp^208:rel^209:rel^210:rel^223:supp Gurven and Kaplan interpret small-scale-society mortality profiles as supporting an adaptive lifespan around 68-78 years and post-reproductive longevity as a robust human feature (lifespan interpretation; post-reproductive longevity). Those demographic interpretations concern ecology, infection, trauma, fertility, social provisioning, and resource transfer, not a controlled ancestral menu.^224:rel

Healthy aging requires preserving function and adequacy as well as reducing risk markers. Acute mycoprotein trials show muscle-protein-synthesis responses in young men (acute response), but they do not establish less frailty in older adults.^225:supp^226:supp^227:supp^228:supp A trial in active older men found that extra protein after exercise and before sleep did not add muscle or strength gain (null supplementation result). Energy reduction, muscle preservation, micronutrient sufficiency, bone support, appetite, and function must be monitored together.^229:rel^230:rel^169:supp^231:rel^232:rel

Cardiometabolic disease

For cardiovascular outcomes, trans fat is more consistently distinguished from broad fat categories: observational synthesis associated trans-fat intake with all-cause mortality (trans-fat association). Saturated-fat interpretation remains conditional on replacement, matrix, and endpoint.^233:supp^234:supp^235:supp^236:supp^237:rel LDL is a relevant pathway, but a surrogate response should not silently become a clinical-event claim.

For type 2 diabetes, evidence supports multiple possible routes.^238:supp^239:supp^240:rel^241:supp^242:supp Across three US cohorts, a healthful plant-based index was inversely associated with incident diabetes after broad adjustment (extreme-decile HR 0.66), while an unhealthy plant-based index was positively associated (HR 1.16) (healthful index; unhealthy index). Vegetarian and vegan interventions report improvements in intermediate outcomes (vegetarian trial), while low-carbohydrate and ketogenic sources report improved insulin sensitivity or glycaemic management (insulin sensitivity; glycaemic management).^34:supp^243:supp^244:rel^245:rel^246:rel These do not prove equivalent long-term complication prevention; they support more than one metabolically useful path under appropriate conditions.

Cancer and neurocognitive evidence is mainly associative.^247:rel^208:rel^248:rel^210:rel^209:rel A composite adherence score based on cancer-prevention recommendations was associated with lower risk of several cancers and lower cancer, circulatory, and respiratory mortality (score association). Higher fish intake was associated with slower memory decline and larger grey-matter volume, while lower omega-3 intake was associated with dementia and cognitive decline (fish association; omega-3 association).^249:supp^248:rel^250:supp^251:rel^252:rel These observations generate hypotheses but remain open to reverse causation, correlated behavior, and measurement error.

Mood, cognition, sleep, digestion, and reported inflammation^253:rel^254:rel^255:rel^256:rel^257:rel

Mood and wellbeing evidence is heterogeneous rather than uniformly positive or negative. A review summarized by a vegan advocacy source found eleven studies associating vegetarian or vegan diets with more depression, seven with less risk or fewer symptoms, and seven with no association (review distribution).^258:supp^259:rel^260:supp^261:rel^262:rel A qualitative ketogenic study reported psychological-wellbeing improvements, while its authors noted that completed randomized trials for depressive symptoms and wellbeing were absent at the time (qualitative finding; trial gap). In an acute inulin comparison, hedonic tone and alertness means were nearly identical between groups (hedonic tone; alertness).^263:supp^264:supp^265:supp^266:supp

Sleep is simultaneously an outcome, mediator, and confounder. A meta-review describes bidirectional relations between poor sleep and mental illness (bidirectional caution).^267:supp^268:supp^269:rel Ketogenic-context evidence flags sleep alteration or deprivation as a possible contributor to mania, hypomania, or psychosis in susceptible people (psychiatric safety signal). Dietary wellbeing claims are weak when sleep, stress, activity, counselling, caffeine, alcohol, and medication change unmeasured.^167:supp^270:rel^271:rel^272:supp

Digestive symptom response is one of the clearest domains of heterogeneity. Low-FODMAP sources report improvement in pain, bloating, gas, constipation, and diarrhea (symptom range), while 50-75% response estimates preserve a 25-50% nonresponder group (response range).^273:supp^274:supp^275:supp^276:supp^277:supp Prolonged restriction may reduce prebiotic exposure, fibre, and bacterial abundance and may increase anxiety, cost, and planning burden (microbiome concern; fibre risk; psychological burden). The evidence supports staged elimination, reintroduction, and a least-restrictive maintenance pattern rather than indefinite restriction (three-phase model).^278:supp^279:rel^280:supp^281:rel^282:contr

Reported inflammation must be separated from measured inflammation. A practitioner quoted in journalism interprets fatigue, digestive problems, mood swings, and mental fog as possible inflammation signals (practitioner interpretation); low-FODMAP material states that symptom improvement does not establish reduced mild gut-lining inflammation (symptom-marker distinction).^283:supp^284:rel^285:supp^286:rel A marker-specific inulin trial in women with type 2 diabetes has greater endpoint precision but limited generalizability (marker trial). Subjective relief and biomarkers should be paired without treating either as intrinsically superior; they answer different questions.^287:supp^218:rel^288:rel

Sub-Domain IV: Individual-Response Architecture

Life stage and sex-related physiology without essentialism

Pregnancy, lactation, growth, aging, menopause, recovery, and activity can change needs and risk, but gender identity or a binary sex label should not substitute for current physiology.^289:supp^290:rel^291:rel^292:supp^102:rel Repository recommendations report about 0.8-0.83 g/kg/day protein for adults, 1.00-1.30 g/kg/day in one older-adult recommendation, and 1.1 g/kg/day during the second and third pregnancy trimesters (adult range; older range; pregnancy). These are institutional estimates for populations, not exact personal requirements.^292:supp^293:supp^294:supp^107:rel^295:supp

A researcher describes loss of ovarian hormone production during menopause as accompanying metabolic changes and possible obesity or metabolic disease (researcher account). The graph links menopause with microbiome, inflammation, hormones, insulin resistance, gastrointestinal motility, and metabolism.^296:supp^297:supp^298:supp^299:supp This supports asking about current hormonal transition and symptoms; it does not support one postmenopausal diet.

Genetics and ancestry limits

Genetic background is one input, not destiny.^300:supp^301:supp^302:rel^303:rel In PREDICT 1, reported genetic contributions were 9.5% for glucose, 0.8% for triglyceride, and 0.2% for C-peptide (reported contributions). Meal composition was the largest reported component in postprandial glucose prediction, followed by genetics, meal context, glycaemic markers, microbiome, and age (predictor ranking).^304:supp^305:rel^306:rel Held-out glucose prediction correlated $R=0.75$ with measurement, while C-peptide prediction was weak and nonsignificant at $R=0.14$ (glucose; C-peptide). Predictability is outcome-specific, and prediction is not demonstrated clinical benefit.^307:rel^304:supp^308:rel^142:contr

Ethnicity, race, ancestry, nationality, and community membership should remain distinct. A cohort found no significant effect modification of plant-diet associations by ethnicity in its sample (null interaction), while authors cautioned that a health-professional sample requires replication across countries and occupations (transport warning).^309:supp^310:supp^311:supp^312:supp^313:rel Community testimony about ancestral foods and “genetic memory” is relevant as cultural interpretation and continuity, not demonstrated genomic mechanism (testimony). Group labels should trigger inquiry, representation audits, or structural analysis, not automatic prescription.

Microbiome and metabolic phenotype

The gut microbiome is described as a dynamic ecosystem involved in digestion, immunity, and neurochemical regulation (definition; function).^314:supp^315:supp^316:supp^317:rel^318:rel Graph exploration links it to dietary pattern, metabolome, obesity, fermented foods, menopause, fibre, insulin resistance, low-FODMAP restriction, and ovarian hormones. That broad connectivity makes it a useful hypothesis and state variable, not a validated standalone prescription engine.^319:rel^320:rel^321:supp^322:rel^211:rel

Microbiome states can be mutable. In a macaque model, maternal diet was associated with offspring microbiome, yet post-weaning diet predicted clustering more strongly in a small subset and partial correction followed dietary change (maternal finding; post-weaning result; partial correction). Animal and small-sample evidence does not authorize deterministic human advice.^323:supp^324:supp^325:supp^326:supp

Activity, disability, conditions, medications, and material life

Activity changes energy use, appetite, glucose handling, and protein demands, while sleep and recovery modify the same outcomes.^327:rel^253:rel^328:rel^329:rel Disability can affect swallowing, digestion, fatigue, mobility, dexterity, sensory or cognitive access, shopping, cooking, and cleanup. Medical conditions can alter tolerance, absorption, texture, timing, and medication interactions.^330:rel^331:rel^332:supp^333:rel Qualitative IBD research documents safe-food scarcity, financial constraints, lost ease of living, inadequate support, and distress around eating out (patient evidence). Such outcomes are not secondary merely because they are absent from a biomarker model.^334:rel^335:supp^307:rel^218:rel^336:rel

A diet is not individualized if it cannot be afforded, accessed, prepared, understood, tolerated, or chosen. Nor is failure to implement an inaccessible plan evidence of weak discipline. Personalization should use the minimum sufficient set of physiological, clinical, functional, cultural, and structural information to make a safe decision, then observe and revise.^337:rel^338:rel^339:rel^340:rel

Sub-Domain V: Lived Evidence and Responsible N-of-1 Learning

Lived reports reveal direction, magnitude, timing, burden, and outcomes that trials often omit. An IBS-D storyteller attributes restored function to strict low-FODMAP eating (personal report); an AIP author reports improved ulcerative-colitis quality of life after about four weeks (personal report); a diarist reports that meal size, timing, combinations, fizzy drinks, fat, and fibre load coincided with symptoms (reported triggers; combination effect).^341:supp^164:rel^276:supp^342:supp^343:supp Another diarist found no triggers after years of tracking (nonresponse), and a dietitian found exclusion overwhelming and unsuccessful (failed attempt).

These reports support real heterogeneity.^344:supp^345:rel^346:rel^347:supp^348:supp Scenario A is that foods, doses, timing, or combinations genuinely change symptoms for some people. Specific temporal patterns, withdrawal/reintroduction, and convergence with structured group evidence increase that scenario's plausibility.^349:rel^350:rel^351:rel^352:rel^353:rel Scenario B is that expectancy, attention, regression from a symptom peak, disease fluctuation, counselling, concurrent changes, stress, sleep, or selection explain some or all of the observed change. IBD is explicitly relapsing-remitting (disease description); a Mediterranean feasibility study bundled diet with personalized counselling while controls received neither (bundled comparison); and an IBD diarist identified sleep loss, work stress, and eating speed as triggers alongside food (contextual triggers). Both scenarios should remain active until observations discriminate them.

Responsible N-of-1 learning is a decision aid, not a miniature universal trial. A prudent process is:

Define one primary target. Name a symptom, function, sleep measure, bowel pattern, glucose response, or other outcome. Do not use “inflammation” when only subjective wellbeing is measured.
Observe baseline variability. Record enough ordinary days to understand fluctuation before changing the diet.
Specify the exposure and replacement. Record foods, dose, timing, preparation, meal size, and what was displaced.
Limit simultaneous changes. A dietitian's first-person warning notes that multiple remedies prevent attribution and can create expensive placebo effects (warning).
Track major co-interventions. Include sleep, stress, activity, medication, supplements, caffeine, alcohol, illness, menstrual or hormonal context, and counselling.
Predefine a review interval and threshold. Decide what magnitude and duration count as benefit, nonresponse, or deterioration before seeing the outcome.
Use withdrawal or reintroduction when safe. Recurrence after reintroduction can strengthen a trigger hypothesis; serious reactions should not be deliberately rechallenged.
Preserve nonresponse and harm. No change should not be relabeled as poor adherence indefinitely.
Broaden after identification. Restriction should move toward the least restrictive tolerable pattern where possible.

Stopping rules are essential. Stop or escalate when there is physical deterioration, dehydration, severe or rapidly worsening symptoms, meaningful weight loss, suspected deficiency, medication-related risk, sleep destabilization, psychiatric symptoms, eating-disorder activation, obsessive tracking, guilt, loss of social function, or unaffordable burden.^354:supp^355:supp^356:supp^357:rel^358:supp A food diary itself can become harmful: one diarist reported obsession, overanalysis, anxiety, and self-punishment (tracking harm; guilt). Restrictive trials in pregnancy, childhood, older age, malnutrition risk, active disease, or eating-disorder history require appropriate professional oversight (risk groups).^355:supp^354:supp^359:supp^360:rel^361:rel

Cultural and collective outcomes require different N-of-1 boundaries. Food sovereignty organizations describe food practice through language, family, land, local economy, treaty rights, and self-determination (community definition; rights framework).^81:supp^18:supp^362:supp^82:rel^84:supp Those effects cannot be tested adequately by an individual biomarker diary. Claims about belonging, continuity, authority, or collective wellbeing require participatory and community-defined outcomes.^8:rel^363:supp^81:supp

Tensions and Debates: Multiple Viable Pathways

The archetypes below are not ranked diets, diagnoses, or endorsements. They are coherent paths supported for some questions and conditions. Each can be well or poorly constructed, and none suits everyone.^364:supp

1. Culturally grounded mixed diet

Core structure. A mixed pattern can combine regionally familiar plant and animal foods, seasonal produce, staple carbohydrates, pulses or grains, selected meat, eggs or dairy, and household preparations. Its strength is flexibility: animal foods can increase protein and micronutrient density, while plant foods can supply fibre, variety, and different matrices. Historical and contemporary evidence supports mixed, adaptive systems more strongly than purity narratives (Edo reconstruction; CHamoru adaptation).

Conditions and monitoring. Monitor adequacy, source-specific saturated fat and replacement, fibre tolerance, processing mix, food safety, affordability, and the balance between convenience and household labor.^61:rel^44:rel^365:supp^366:supp The pattern can accommodate religious observance and community continuity.

Tradeoffs and gaps. “Mixed” can conceal high processed-meat or refined-food intake just as easily as diversity. Cohort categories compress fish, yogurt, eggs, and processed meat into broad animal exposure; long-term head-to-head trials of culturally specific mixed patterns are sparse.^37:rel^367:rel^39:rel^368:rel

2. High-quality plant-forward path

Core structure. This path emphasizes pulses, soy foods, whole grains, nuts, seeds, vegetables, fruits, and culturally appropriate preparations, with optional or no animal foods. The strongest disease signal distinguishes healthful from unhealthful plant foods: the diabetes associations reverse by plant-food quality (healthful; unhealthful). Plant sources can provide substantial protein, and soy or potato can score highly for amino-acid quality (soy; potato).^369:supp^370:supp^108:supp^371:supp^372:supp

Conditions and monitoring. Ensure sufficient energy and protein volume; monitor B12 where animal foods are absent, iron, zinc, calcium, iodine, vitamin D, omega-3 strategy, and tolerance of fibre or legumes. Preparation and fermentation may improve usability.^117:supp^373:supp^108:supp^105:supp^106:supp Life stage, low appetite, pregnancy, and older age can increase the importance of density and monitoring.

Tradeoffs and gaps. Plant origin does not guarantee quality; refined products and analogues differ from pulses and tofu.^374:supp^375:supp^376:supp^28:supp^377:supp Observational mortality and diabetes evidence remains vulnerable to confounding. Affordability claims for pulses omit fuel, time, water, and digestive tolerance.^378:rel^379:rel Mood evidence for vegan or vegetarian patterns is directionally heterogeneous (review distribution).

3. Lower-carbohydrate or metabolic-control path^380:contr^137:supp^381:supp^108:supp^382:contr

Core structure. This path reduces carbohydrate to target glycaemia, insulin resistance, satiety, or selected symptoms and replaces it with specified proteins, fats, and lower-carbohydrate plant foods. Intervention sources report improved insulin sensitivity or glycaemic management in some people (insulin sensitivity; management). Qualitative ketogenic reports describe improved wellbeing for some participants (experience).^151:supp^383:rel^384:rel^160:supp^85:supp

Conditions and monitoring. Specify replacement source; track glucose where relevant, lipids, bowel function, hydration, sleep, energy, nutrient adequacy, medication requirements, and psychiatric vulnerability. A plant-rich lower-carbohydrate pattern is not the same exposure as one dominated by processed meat and animal fat.^385:rel^386:rel^387:supp

Tradeoffs and gaps. Early fatigue, headache, dizziness, and constipation are reported (adaptation symptoms). Cohort mortality associations move in opposite directions by replacement source (plant; animal).^356:supp^388:rel^389:rel^390:supp Sleep destabilization can be a serious boundary in susceptible people. Long-term complication and mortality evidence is insufficient to treat short-term glycaemic improvement as universal superiority.^391:supp^392:rel^393:rel^394:supp^395:supp

4. Aquatic or pescatarian path

Core structure. This path uses fish, shellfish, seaweed, algae, or other locally appropriate aquatic foods as important protein and micronutrient sources, usually alongside plants.^356:supp^396:supp^397:supp^267:supp^398:rel Aquatic foods can bundle protein, iodine, selenium, vitamins, and long-chain omega-3 fats (co-nutrient account). Cohort sources distinguish fish from red and processed meat in reported cardiovascular associations (source-specific cohort).

Conditions and monitoring. Consider species, mercury or other contaminants, pregnancy and childhood guidance, pathogen control, allergy, seasonality, refrigeration, cost, harvest rights, and ecological system. In subsistence communities, aquatic mammals or fish can carry cultural and nutritional importance not captured by generic seafood guidance (subsistence context).^123:rel^120:supp^399:supp^400:rel^401:supp

Tradeoffs and gaps. Aquatic foods range from wild capture to aquaculture, algae, and shellfish; category averages obscure ecological and nutritional variation. Affordability claims conflict and need household-level comparison (outreach claim; cost deterrent).^402:supp^403:rel^404:rel^405:supp^406:rel

5. Fermented or traditional-preparation path

Core structure. This path centers culturally established fermenting, soaking, cooking, drying, pickling, or other transformations.^407:supp^408:rel^123:rel Its strongest support concerns preservation, safety under controlled conditions, cultural continuity, and selected nutrient-access effects (preservation; safety; mineral mechanism). Tempeh illustrates that fermentation is also labor, cottage production, wrapping material, perishability, and cultural identity (traditional process; production context).^409:rel^410:rel^411:rel^412:rel^413:rel

Conditions and monitoring. Identify substrate, starter or microbial community, salt and sugar, heat treatment, storage, contamination risk, tolerance, and whether live organisms remain. Preserve cultural authority and do not rebrand established foods as placeless innovations.

Tradeoffs and gaps. “Fermented” is too broad for a clinical claim.^414:supp^198:supp^415:supp^416:rel^417:rel Observational microbiome associations and small mood studies do not establish a category-wide benefit (microbiome association; preliminary mood observation). Standardization may improve safety while narrowing flavor, microbial diversity, producer control, or local knowledge (producer concern).^418:supp^419:supp^420:rel^421:rel^422:supp

6. Medically constrained or symptom-targeted path

Core structure. This path uses a defined restriction or modification for a diagnosed condition, symptom, swallowing need, allergy, intolerance, renal context, metabolic target, or medication interaction.^8:rel^75:rel^423:rel^424:rel^416:rel Low-FODMAP practice is the clearest example of a staged intervention with elimination, reintroduction, and individualized maintenance (three phases). Texture modification, fortified foods, supplements, or high-density sources may be necessary in other contexts.^200:supp^201:supp^425:supp^426:contr^199:supp

Conditions and monitoring. Define the target, baseline, duration, adequacy plan, medication context, professional role, response threshold, and route to broaden the diet. Pair symptoms with clinical indicators when disease activity matters. Screen eating-disorder and malnutrition risk.^427:supp^428:supp^429:supp^430:rel

Tradeoffs and gaps. Restriction can reduce fibre, microbial diversity, affordability, pleasure, and social participation and can increase anxiety or guilt (microbiome concern; social isolation). Symptom control is not necessarily disease modification.^278:supp^280:supp^281:rel^431:rel^432:rel Nonresponse is common enough to require stopping rules rather than escalating restriction indefinitely.

A Unified Multidimensional Feeding Framework^390:supp^433:rel^434:rel

The axes

A feeding decision should be located on at least twelve axes:^435:rel^436:supp^231:rel^437:supp^438:rel

Axis	Decision question
Goal and endpoint	Is the target adequacy, growth, function, glycaemia, lipids, bowel symptoms, mood, sleep, disease events, quality of life, cultural continuity, or food security?
Evidence distance	Is support material, compositional, mechanistic, observational, interventional, clinical-event, community, practitioner, or first-person evidence?
Explicit substitution	What food, nutrient, labor practice, or social activity decreases when another increases?
Nutritional architecture	Are energy, protein, indispensable amino acids, fats, carbohydrate quality, fibre, micronutrients, and hydration adequate together?
Matrix and transformation	Is food intact, ground, cooked, fermented, extruded, fortified, or reassembled, and what does that change?
Physiological and clinical state	What current life stage, hormones, function, conditions, allergies, medications, and malnutrition risks matter?
Observed response	What changes, in which direction, after what latency, and with what reproducibility?
Access and labor	Can food be afforded, obtained, stored, prepared, eaten, and cleaned up with available money, time, fuel, equipment, transport, and assistance?
Culture and governance	Does the path support religion, identity, pleasure, family participation, sovereignty, and legitimate control?
Durability and burden	Can it persist without unacceptable restriction, anxiety, isolation, caregiver load, or opportunity cost?
Production system	What land, water, energy, biodiversity, labor, ownership, and supply-chain consequences apply locally?
Uncertainty and reversibility	How strong is the evidence, what is the cost of error, and can the decision be safely revised?

Practical decision process^439:supp^440:rel^218:rel

Name the decision, not a diet identity. Start with a concrete outcome and time horizon.
Identify the binding constraint. It may be nutrient density, glycaemia, symptoms, appetite, money, housing, disability access, cultural disruption, or preparation capacity.
Define the feasible choice set. Exclude options that are unsafe, inaccessible, culturally unacceptable, or incompatible with necessary care.
Describe the proposed substitution. “Eat less fat,” “eat more plants,” or “avoid processing” is incomplete without naming replacements and forms.
Protect adequacy before optimization. Check energy, protein, relevant micronutrients, hydration, and life-stage needs, especially under restriction.
Match evidence to the endpoint. Do not use an acute amino-acid study to promise longevity, a testimonial to estimate prevalence, or an official threshold as an individual optimum.
Choose the least burdensome plausible path. Favor familiar, affordable, reversible changes before high-complexity restriction or testing when expected benefit is similar.
Measure a small set of meaningful outcomes. Include benefits, harms, function, burden, and adherence. Do not optimize a convenient biomarker while quality of life deteriorates.
Review at a predefined interval. Continue when benefit is meaningful and burden acceptable; adapt when response is mixed; stop when there is no meaningful benefit or harm emerges.
Escalate appropriately. Severe symptoms, suspected deficiency, major weight change, pregnancy complications, medication interactions, active disease, psychiatric destabilization, swallowing problems, or malnutrition risk require qualified care rather than further self-experimentation.

This process is useful precisely because it does not assume that every person needs maximal data. PREDICT-style biomarker prediction may help selected decisions, but the graph shows personalized nutrition as a comparatively narrow biomarker cluster while microbiome, IBD, food insecurity, and lived evidence connect to broader physiological and structural variables. Agent hypothesis, not asserted by any source: good personalization is not maximal measurement; it is the minimum sufficient information that changes a safe, feasible decision, followed by observation and revision. This hypothesis is grounded in outcome-specific prediction (glucose; C-peptide) and the clinical priority of food and housing provision (food referral; housing).^441:rel^14:supp^11:rel

Evidence Posture and Stakeholder Incentive Matrix

Evidence posture

Evidence type	What it can establish well	What it cannot establish alone	Key audit
Archaeology/material traces	Contact, residues, tissue signatures, tools, documented objects	Complete menus, prevalence, modern optimality	Proxy window, baseline, contamination, triangulation
Composition/digestion	Nutrient presence, amino-acid profile, release, acute absorption	Long-term disease, feasibility, cultural fit	Product, preparation, assay, serving, host
Mechanism/biomarker	Plausible pathway and intermediate response	Net clinical benefit or lifespan	Validation, competing pathways, endpoint relevance
Observational cohort	Long-term associations and heterogeneity	Guaranteed counterfactual causality	Exposure error, substitution, confounding, reverse causation
Randomized intervention	Causal package effect during observed assignment	Automatic lifetime or class-wide effect	Comparator, adherence, attrition, duration, harms
Systematic review	Structured synthesis under explicit criteria	Repair of biased or heterogeneous primary evidence	Protocol, search, deviations, missing outcomes, overlap
Guideline/regulation	Institutional judgment, process, thresholds, legal status	Individual optimum or comparative causal certainty	Decision trail, evidence-to-wording map, conflicts
Community/cultural knowledge	Meaning, practice, priorities, local mechanisms, collective outcomes	Universal biochemical effect without appropriate testing	Authority, ownership, context, translation boundary
Practitioner/journalism	Implementation detail, candidate mechanisms, real-world bundles	Prevalence or isolated causality	Selection, co-interventions, service and narrative incentives
First-person report	Experienced direction, timing, burden, function	Population frequency or clean causal allocation	Baseline, challenge, co-interventions, fluctuation, selection

Confidence should be described as convergent, provisionally supported, contested but discriminable, underdetermined, or procedurally opaque.^442:rel^443:rel^444:rel^445:rel^446:rel Repetition by an agency, journal, company, advocacy organization, or media network increases reach, not evidentiary weight. A well-evidenced claim has a defined comparator and estimand, suitable design, validated measurement, explicit assumptions, complete outcome and funding disclosure, result-conclusion concordance, and convergence across differently interested teams.

Stakeholder incentives

Stakeholder	Legitimate contribution	Incentives or blind spots to inspect	Evidence that would strengthen trust
Governments/regulators	Surveillance, standards, public programs, legal review	Administrability, political mandates, jurisdiction, agricultural and health pressures, policy continuity	Member-level disclosures, evidence-to-decision trail, implementation audits
Academic researchers/journals	Measurement, causal designs, synthesis, critique	Funding, publication, novelty, citations, paradigm and career commitments	Preregistration, full outcomes, data/code, adversarial replication
Food and technology industries	Product data, scale, technical expertise, supply innovation	Market growth, favorable comparators, liability, proprietary control, delayed or selective publication	Independent analysis, publication rights, full sponsor role, cross-sponsor replication
Professional bodies/practitioners	Clinical feasibility, safety, implementation, case patterns	Service models, authority, member interests, memorable successes, referral selection	Outcome registries, harms and nonresponse, clear scope, external evaluation
Advocacy organizations/critics	Neglected harms, accountability, community mobilization, document analysis	Donor alignment, ideological coherence, opponent-only scrutiny, attention incentives	Primary records, own-interest disclosure, symmetric standards, correction policy
Journalists/platforms	Public access, record linkage, ecological detail, first-person visibility	Novelty, conflict, dramatic narratives, editorial selection	Source links, uncertainty, denominator, corrections, distinction between result and interpretation
Communities/Indigenous organizations	Local knowledge, cultural authority, governance, outcomes outsiders omit	Resource and recognition needs, strategic presentation of stable tradition	Community governance, internal plurality, transparent aims, long-term participatory evaluation
Storytellers/participants	Timing, lived burden, function, practical signals	Recall, expectancy, coherence, identity, publication selection	Prospective records, null and harm periods, challenge/rechallenge where safe

The corpus provides empirical reason for heightened sponsorship scrutiny without treating funding as a verdict.^11:rel^14:supp^20:rel^364:supp^447:supp A Cochrane review reported 863 favorable conclusions per 1,000 sponsored studies versus 644 per 1,000 controls and 638 versus 502 favorable efficacy results (conclusions; results). Reviews of artificially sweetened beverages reported undisclosed author conflicts in 42% and more favorable conclusions among conflicted authors, though estimates were imprecise (undisclosed conflicts; association).^352:rel^351:rel^349:rel^350:rel^353:rel These patterns support scrutiny of control, disclosure, and framing; they do not prove each sponsored conclusion false.

The evidence also documents selective disclosure where guideline conflicts were aggregated rather than attributed to individual members and monetary value was omitted (aggregate disclosure; value omission). It documents strategic framing at the portfolio level where conclusions favor sponsors more often than results and where result-conclusion discordance is reported (discordance).^448:supp^449:rel^450:rel^451:rel Government procedures also include ethics review, multiple evidence modes, public meetings, and comment channels (review modes; ethics vetting; public access). Formal controls and influence channels coexist; neither should erase the other.^452:supp^453:supp^454:rel^455:supp^456:supp

Symmetric skepticism applies to critics. Advocacy allegations about commodity-linked appointments and testimony identify testable pathways but are not adjudicated violations (advocacy allegation).^457:rel^458:rel^455:supp^452:supp^459:supp A strong critique discloses its own commitments, links primary records, applies standards to allies, distinguishes allegation from finding, and states what would change its conclusion.

Honest Assessment: Convergence, Uncertainty, and Discriminating Evidence^460:contr^461:rel^462:supp^463:rel^464:rel

Strongest convergence

Dietary plurality is historically and contemporaneously real. Different proxies, archives, foodways, and modern patterns show adaptation across ecology, mobility, culture, and technology; none establishes one authentic human menu (methods; foodways).
Food source, matrix, preparation, and substitution matter. Protein scores vary within kingdoms; nut structure changes absorbed energy; fermentation changes selected mineral availability; carbohydrate and fat associations change with replacement (protein variation; matrix; fermentation; substitution).
Quality within a named category can reverse associations. Healthful versus unhealthful plant indices and plant versus animal low-carbohydrate replacements show that labels alone are inadequate (plant quality; replacement direction).
Short-term efficacy and durable benefit differ. Named-diet effects commonly diminish by twelve months, while restriction and adherence burdens influence net value (durability).
Response distributions include benefit, no response, and harm. Digestive evidence and lived reports preserve all three rather than supporting a universal responder story (response range; failed attempt; tracking harm).
Access, culture, and control are causal parts of implementation. Food, housing, transport, disability access, religious fit, time, equipment, and sovereignty determine whether nutritional advice can become practice (food referral; disability access; sovereignty).
Evidence authority depends on traceability, not prestige. Design, comparator, measurement, funding control, synthesis, and translation must remain visible; sponsor-related conclusion patterns and incomplete disclosures justify scrutiny across institutions (sponsorship pattern; disclosure issue).

Live uncertainties^465:supp^466:supp^467:supp^468:supp

It remains unclear which explicit whole-food substitutions produce durable clinical-event benefits across cultures and life stages. Long-term mortality evidence is mostly observational; randomized evidence is stronger for short-term biomarkers and weight. It remains unclear how much saturated-fat risk differs by food matrix beyond replacement effects; which UPF features drive package-level intake; which fermented products benefit which hosts; and whether microbiome-guided advice improves patient-valued outcomes beyond careful conventional support.^469:rel^20:rel^21:rel^470:rel^465:supp

It also remains unclear how to compare well-constructed plant-forward, mixed, aquatic, and lower-carbohydrate paths when energy, food quality, support, cost, and adherence are genuinely matched. Existing studies often compare unequal packages. Emerging proteins need independent product-specific evidence on allergenicity, gastrointestinal tolerance, contaminants, chronic intake, affordability, labor, ownership, and life-cycle effects. Ancestry-based individual prescription remains especially underdetermined.^233:supp^234:supp^368:rel^222:supp^471:rel

Mood, cognition, and broad “inflammation” claims remain weaker than digestive-symptom evidence. Qualitative ketogenic experience is meaningful but causally open; vegan/depression evidence is heterogeneous; acute prebiotic contrasts include null-like outcomes; and subjective inflammation language often lacks markers.^207:rel^234:supp^233:supp^472:supp^425:supp Harms reporting is incomplete: one systematic review reported a median 54% of studies failing to report harms data (harms under-reporting).

What would discriminate scenarios^473:supp^474:rel^475:rel^476:supp^477:rel

Long-duration randomized substitution trials comparing specific foods and matrices, with objective adherence, clinical events, function, quality of life, harms, and sufficient diversity.
Pragmatic trials matching contact, counselling, cost, convenience, processing, and energy across multiple viable dietary paths.
Product-specific fermentation and processing experiments that characterize substrate, organisms, structure, eating rate, nutrient release, additives, and host response.
Prospective N-of-1 and subgroup studies retaining responders, nonresponders, harmed participants, and dropouts, with predefined outcomes and safe withdrawal/rechallenge.
Precision-nutrition trials comparing biological targeting, attentive standard care, structural support, and combined support on patient-valued outcomes, burden, cost, and equity.
Participatory foodway research measuring sufficiency, health, labor, ecological resilience, cultural continuity, and decision authority together.
Governance audits linking member-level interests, amounts, dates, recusals, assignments, votes, evidence exclusions, public comments, agency edits, and final wording.
Independent adversarial replication across teams with different financial and ideological interests, including full protocols, data, harms, and null outcomes.

Widely repeated positions versus well-evidenced claims^337:rel

“Saturated fat is harmful,” “plant-based is healthy,” “low carbohydrate is dangerous,” “low carbohydrate reverses metabolic disease,” “UPFs are harmful,” “fermented foods heal the gut,” “ancestral eating is optimal,” and “personalized nutrition finds your ideal diet” are widely repeated positions. Their strongest evidence-supported forms are narrower: specified saturated-fat substitutions alter LDL and may alter events; plant-food quality matters; lower-carbohydrate metabolic effects depend on replacement and person; one tested ultra-processed package increased intake; specified fermentations improve preservation or selected mineral access; archaeology documents plurality rather than optimality; and selected biomarkers can be predicted better than others.^478:supp^479:rel^480:supp^481:rel^482:supp Narrowing is not evasion. It is the difference between repetition and a claim whose evidence can be audited.

Significance^483:rel^484:rel^222:supp^14:supp

Human alimentation connects molecular nutrition to ecology, history, family, labor, disability, medicine, markets, and political control. This synthesis matters because it replaces two equally limiting models: the universal diet and the anything-goes relativism.^207:rel^485:contr^11:rel^486:rel^487:rel The evidence does not say every diet is equally suitable. It says suitability is conditional and can be investigated.

The central practical unit is a feasible food substitution sustained over time, embedded in a dietary episode and food system.^352:rel^351:rel^349:rel^350:rel^353:rel This is an agent synthesis, not a direct source claim. It is supported by opposite mortality directions for different replacement sources (plant; animal), quality reversal within plant patterns (quality reversal), fading named-diet effects (durability), bundled counselling and dietary interventions (bundle), and material-access constraints (clinical provision). Testing it requires studies designed around actual substitutions, support, access, and durability rather than diet identities alone.^38:rel^488:rel^15:supp^12:supp

A multidimensional feeding framework therefore supports plural paths with disciplined boundaries. It values cultural mixed diets, high-quality plant-forward patterns, lower-carbohydrate metabolic strategies, aquatic patterns, traditional fermentations, and medically constrained approaches when their conditions are met. It refuses to infer individual destiny from sex, gender, ancestry, genotype, disability, or microbiome alone.^368:rel^489:rel^471:rel It treats anecdotes as attributed signals, not votes. It asks whose labor and resources make a recommendation possible. It audits institutions, companies, advocates, researchers, and critics with the same questions.^368:rel^475:rel^484:rel And it expects a diet to remain revisable when life stage, health, access, goals, or evidence changes.

Provenance and Method Note^488:rel^1:rel^490:supp^491:rel^492:rel

This meta-synthesis integrates nine complete scope syntheses produced from approximately 239 source submissions across nine investigations. Submissions are not the same as unique, successfully parsed, evidence-bearing sources: duplicate URLs, retrieval failures, parsing differences, decomposition, and materialization mean submitted counts, materialized source counts, and usable facts can differ. The work therefore does not claim completion of the original 400-source target and does not inflate submission count into successfully parsed evidence.

Source gathering intentionally maintained a non-academic majority. Community organizations, Indigenous sources, public testimony, patient narratives, practitioner accounts, journalism, museums, archives, advocacy, government, industry, and academic research were used for the questions each could answer. Non-academic material is integrated throughout access, foodways, preparation, symptom learning, cultural healing, governance, and feasibility rather than confined to an anecdote appendix. Source neutrality here means equal scrutiny, not pretending all methods establish the same thing.^493:rel^494:rel^495:rel

The synthesis process read all nine complete reports and preserved their existing fact and concept links. It also used organic graph exploration across key concepts from every scope. That walk identified cross-scope bridges among protein and fibre/fats/iron/mortality; food insecurity and housing/healthcare/employment; foodways and gender/climate/oral history; fermentation and microbial/aquatic/carbohydrate systems; microbiome and menopause/fibre/low-FODMAP restriction; IBD and food diaries/eating disorders/social burden; and systematic reviews and conflicts/selective reporting/guidelines. New cross-scope interpretations were labeled as agent hypotheses rather than attributed to sources.

Scope 9 carries one operational caveat.^496:rel^101:rel^497:rel^75:rel At synthesis time its task ledger showed 248 completed jobs, including 32 completed synthesize_concept jobs, and one persistently running extract_concepts job that the user identified as a development-system bug. That scope proceeded on substantial materialized evidence, but it was not a formally clean drain; graph completeness may therefore be affected at the margin.^494:rel^498:rel^499:rel

Finally, repository source count measures provenance duplication, not underlying study size or independent replication. Many facts have one ingested source even when that source is a review. Confidence in this report follows design, directness, specificity, transparency, independent convergence, and relevance to the decision, not repository status, institutional prestige, or frequency of repetition.