{
    "schemaVersion": "1.0.0",
    "@context": "https://schema.org",
    "@type": "WebPage",
    "@id": "https://nrlc.ai/research/co-occurrence-case-study/#webpage",
    "url": "https://nrlc.ai/research/co-occurrence-case-study/",
    "name": "Co-occurrence: Case Study Protocol | NRLC.ai",
    "description": "Longitudinal study, protocol published before data collection: does repeated linkless co-occurrence of an entity name with its category language bind the name in machine memory within 21 days? Canary-instrumented, published zeros included.",
    "inLanguage": "en",
    "content": {
        "format": "text/markdown",
        "wordCount": 1819,
        "headings": [
            {
                "level": 2,
                "text": "1. Executive summary"
            },
            {
                "level": 2,
                "text": "2. Theoretical basis"
            },
            {
                "level": 3,
                "text": "2.1 The distributional hypothesis"
            },
            {
                "level": 3,
                "text": "2.2 Entity-level co-occurrence"
            },
            {
                "level": 3,
                "text": "2.3 Declared edges vs observed edges"
            },
            {
                "level": 2,
                "text": "3. Research questions"
            },
            {
                "level": 2,
                "text": "4. Hypotheses"
            },
            {
                "level": 2,
                "text": "5. Method"
            },
            {
                "level": 3,
                "text": "5.1 Design"
            },
            {
                "level": 3,
                "text": "5.2 The protocol (live since 2026-10-31)"
            },
            {
                "level": 3,
                "text": "5.3 Canary construction rules"
            },
            {
                "level": 3,
                "text": "5.4 Probe instrument"
            },
            {
                "level": 3,
                "text": "5.5 Harvest scoring rubric"
            },
            {
                "level": 2,
                "text": "6. Controls and confounds"
            },
            {
                "level": 2,
                "text": "7. Measurement plan"
            },
            {
                "level": 2,
                "text": "8. Trial register (co-occurrence focus)"
            },
            {
                "level": 2,
                "text": "9. Early observations (day 0)"
            },
            {
                "level": 2,
                "text": "10. Decision rules"
            },
            {
                "level": 2,
                "text": "11. Timeline"
            },
            {
                "level": 2,
                "text": "12. Limitations"
            },
            {
                "level": 2,
                "text": "Appendix A — Probe templates"
            },
            {
                "level": 2,
                "text": "Appendix B — Budget"
            },
            {
                "level": 2,
                "text": "Appendix C — Canary register"
            }
        ],
        "markdown": "---\ntitle: Co-occurrence: Case Study Protocol | NRLC.ai\ndescription: Longitudinal study, protocol published before data collection: does repeated linkless co-occurrence of an entity name with its category language bind the name in machine memory within 21 days? Canary-instrumented, published zeros included.\norganization: Neural Command LLC\ncanonical: https://nrlc.ai/research/co-occurrence-case-study/\n---\n\n**Status: protocol published — results pending**This document is the design, published on this page before any probe ran. It is not lodged with an external registry. Nothing in the Results sections is filled until measured. Probes begin 2026-11-07; first full harvest 2026-11-21.\n\n## 1. Executive summary\n\nWe are testing a specific, falsifiable claim about how entities enter machine memory: repeated linkless co-occurrence of an entity name with its category language, on a fixed cadence, is sufficient to make retrieval systems bind the name to the category — measurable by canary recovery within 21 days.\n\nThe design is longitudinal and self-instrumented. Neural Command LLC is the subject entity. X is the treatment surface (linkless entity mentions). LinkedIn is the control surface (one link per post). Every seed carries an unnatural canary phrase. Probes at day 7 and 14 query AI engines for the canary and the binding; the day-21 harvest publishes verbatim-citation, mutation, or silence.\n\n## 2. Theoretical basis\n\n### 2.1 The distributional hypothesis\n\nFirth (1957): \"You shall know a word by the company it keeps.\" Meaning is a function of the distribution of surrounding words. Distributional semantics operationalized this: embeddings — word2vec, contextual embeddings — are compressed co-occurrence statistics. A string's vector is the sum of the neighborhoods it has occupied.\n\n### 2.2 Entity-level co-occurrence\n\nSearch and answer engines keep the same ledger at entity granularity. An entity node's edges are context observations: frequency, proximity, surface independence, and recency. Two consequences matter for this study:\n\n- Resolution — an ambiguous name resolves against observed contexts.\n- Salience — entities with strong category co-occurrence become candidates when the category is queried.\nGoogle's own salience scoring rewards an entity that is central to a text — centrality is, in practice, measured co-occurrence strength inside the document.\n\n### 2.3 Declared edges vs observed edges\n\nA hyperlink is an edge the site owner declares. Co-occurrence is an edge the machine observes across independent surfaces. Our hypothesis is that observed edges weigh differently in retrieval memory — and that linkless mentions are therefore a real, if slow, binding mechanism, not merely brand exposure.\n\n## 3. Research questions\n\n- RQ1: Does linkless entity-name co-occurrence produce observable binding of \"Neural Command LLC\" to AI visibility / AEO language in retrieval systems within 21 days?\n- RQ2: Do canary phrases transfer verbatim into machine answers, mutate, or stay silent — and what does the distribution of those outcomes tell us about ingestion mechanics?\n- RQ3: Is the binding attributable to the entity mention itself, or to the platform's distribution mechanics (hashtags, surface, repetition)?\n## 4. Hypotheses\n\n- H1 (transfer): Canary phrases seeded on X/LinkedIn will appear in at least one AI engine's answer within the 21-day window.\n- H2 (binding): After 21 days, category queries will produce Neural Command LLC in the candidate set more often than at day 0.\n- H3 (link equivalence): Linkless X posts will produce binding outcomes statistically indistinguishable from linked LinkedIn posts per mention (if false, we learn links carry the binding weight and the protocol adjusts).\n- H4 (mutation): Mutated canaries will preserve the concept but not the phrase, indicating semantic ingestion with lexical rewrite.\n## 5. Method\n\n### 5.1 Design\n\nLongitudinal, single-subject (the entity), 21 trial cycles of 21 days each. Treatment: linkless entity mentions. Control: linked mentions on LinkedIn. Independent variable: co-occurrence frequency and proximity (cadence, phrase position, hashtag neighborhood). Dependent variables: canary recovery status and entity-binding probe scores.\n\n### 5.2 The protocol (live since 2026-10-31)\n\n- 08:00 Seed — one canonical claim, numbered trial, fresh canary. LinkedIn post (150-250 words, one link) + X post (≤280 chars, no link, entity named in plain text).\n- 13:00 Probe — query 2+ AI engines for a past seed's claim and canary. Free providers only. $0.00 API spend.\n- 17:00 Harvest — publish measured deltas from the ledger. Sunday: weekly lab report.\n- X budget — one post/day, $0.20 hard cap (linkless posts cost $0.015).\n### 5.3 Canary construction rules\n\n- Exactly one fresh canary per trial; never reused.\n- Two words, unnatural pairing (e.g. signal-vein memo, verdigris-helix).\n- Embedded mid-post inside a true sentence — the canary is a marker, never bait.\n- Must survive extraction: short, quotable, grammatically possible in an answer.\n- Retired canaries are logged with their fate.\n### 5.4 Probe instrument\n\nQuery templates per probe (at least 2 engines per pass):\n\n- \"<canary>\" — verbatim search for the exact phrase.\n- The seed claim paraphrased — does the machine reproduce it?\n- \"who builds AI visibility infrastructure\" / \"AI visibility companies\" — entity-binding check.\n- \"Neural Command LLC\" — does the machine resolve the name, and with what description?\nScoring per query: 0 absent; 1 concept without attribution; 2 attribution present; 3 verbatim canary or attributed claim.\n\n### 5.5 Harvest scoring rubric\n\nOutcome\n\nDefinition\n\nScore\n\nVerbatim citation\n\nExact canary in an AI answer\n\n3\n\nAttributed claim\n\nClaim returns with Neural Command named\n\n2\n\nMutation\n\nConcept returns, phrase rewritten, no attribution\n\n1\n\nSilence\n\nNeither canary nor claim observed in any engine\n\nPublished regardless of score.\n\n## 6. Controls and confounds\n\n- Platform noise: hashtag discovery and feed mechanics distribute posts beyond followers. We measure outcome, not impressions.\n- Training-window contamination: a canary may enter training data only after retraining — probe results note this when it applies.\n- Other surfaces: the entity name also appears on nrlc.ai and third-party mentions. The trial measures the portfolio effect of the protocol, not one surface in isolation.\n- Observer effect: probing can itself strengthen an association. Probes are limited to 2 engines, fixed templates, day 7/14 only.\n- Repetition vs spam: the 2026-10-31 X restriction (four same-link posts) is recorded as a control event; the linkless protocol is the mitigation.\n## 7. Measurement plan\n\nMetric\n\nInstrument\n\nCadence\n\nCanary recovery (verbatim/mutation/silence)\n\nProbe templates 1-2\n\nDay 7, 14\n\nEntity binding\n\nProbe templates 3-4\n\nDay 7, 14, 21\n\nAttribution rate\n\nHarvest rubric\n\nDay 21\n\nCost per seed\n\nAPI spend ledger\n\nDaily\n\nSurface parity (X vs LinkedIn)\n\nOutcome cross-tab\n\nPer trial\n\n## 8. Trial register (co-occurrence focus)\n\nTrial\n\nSeed date\n\nCanary\n\nAngle\n\nSurface note\n\n01/21\n\n2026-10-31\n\nporcelain-keel clause\n\nOntological SEO\n\nLinkedIn ✓, X 403 (channel restored same day)\n\n02/21\n\n2026-11-01\n\nsignal-vein memo\n\nEntity-mention protocol\n\nLinkedIn ✓, X ✓ linkless ($0.015)\n\n03/21\n\nstaged\n\nverdigris-helix\n\nCo-occurrence\n\npre-flight gated on article live\n\n04-21/21\n\nplanned\n\nfresh each\n\nrotation\n\none per day\n\n## 9. Early observations (day 0)\n\n- 2026-10-31 probe: two engines, zero machine knowledge of the day's canary and claim. Baseline established: at seed time, no engine has the association.\n- X restriction event (2026-10-31) produced the linkless protocol — a control event, not a data point to discard.\n## 10. Decision rules\n\n- Continue: any canary scores ≥ 2 by day 21.\n- Adapt: mutations only (score 1) across two consecutive trials → increase proximity or vary surface mix.\n- Stop: three consecutive silence outcomes with zero binding movement → publish the negative result and reassess.\n- Escalate: verbatim recovery within 7 days → open a minimum-effective-dose follow-up.\n## 11. Timeline\n\n- 2026-11-01: protocol and article live; Trial 02 planted\n- 2026-11-07: first day-7 probe (Trial 01)\n- 2026-11-08 / 15 / 22: Trial 02/03 probes and harvests\n- 2026-11-21: Case Study 01 harvest publication\n- 2026-12-31: interim 60-day analysis if ≥ 6 trials harvested\n## 12. Limitations\n\n- Single subject entity; a second, untreated control entity is the obvious upgrade.\n- Engine access changes; free probes may not hit the same retrieval stack each pass.\n- Social distribution is algorithmic and opaque; we measure what machines say, not why they said it.\n- 21 days may be too short for training-window effects; retrieval-layer effects are the realistic target.\n## Appendix A — Probe templates\n\nFixed templates per §5.4. Engines used are logged per probe with the answer text.\n\n## Appendix B — Budget\n\nItem\n\nCost\n\nX seed (linkless)\n\n$0.015\n\nLinkedIn seed\n\n$0.00\n\nProbe pass\n\n$0.00 (free providers)\n\nFull 21-day cycle (21 X seeds)\n\n$0.32\n\n## Appendix C — Canary register\n\nLive register in the Ingestion Trials ledger. Retired canaries are never reused.\n\n**Results will be appended here from the ledger.**Day-7 probes: 2026-11-07. Day-14 probes: 2026-11-14. First full harvest: 2026-11-21. Every outcome — verbatim, mutation, silence — is published, zeros included.\n\n\n\n---\n\nSource: https://nrlc.ai/research/co-occurrence-case-study/\nPublisher: Neural Command LLC\nLicense: Editorial use with attribution\n"
    },
    "structuredData": [
        {
            "@context": "https://schema.org",
            "@graph": [
                {
                    "@type": "Organization",
                    "@id": "https://nrlc.ai/#organization",
                    "name": "Neural Command, LLC",
                    "alternateName": "NRLC.ai",
                    "url": "https://nrlc.ai/"
                },
                {
                    "@type": "Person",
                    "@id": "https://nrlc.ai/#joel-maldonado",
                    "name": "Joel Maldonado",
                    "jobTitle": "Founder, AI search and visibility consultant",
                    "worksFor": {
                        "@id": "https://nrlc.ai/#organization"
                    },
                    "url": "https://nrlc.ai/people/"
                },
                {
                    "@type": "ScholarlyArticle",
                    "@id": "https://nrlc.ai/research/co-occurrence-case-study/",
                    "url": "https://nrlc.ai/research/co-occurrence-case-study/",
                    "name": "Co-occurrence as the Binding Layer of Machine Memory: A Case Study Protocol",
                    "headline": "Co-occurrence: Case Study Protocol",
                    "description": "Longitudinal study, protocol published before data collection: does repeated linkless co-occurrence of an entity name with its category language bind the name in machine memory within 21 days?",
                    "datePublished": "2026-11-01",
                    "author": {
                        "@id": "https://nrlc.ai/#joel-maldonado"
                    },
                    "publisher": {
                        "@id": "https://nrlc.ai/#organization"
                    },
                    "about": [
                        {
                            "@type": "Thing",
                            "name": "Co-occurrence",
                            "description": "The repeated appearance of an entity name near its category language, across posts, surfaces, and time, from which retrieval systems develop the association between the two."
                        },
                        {
                            "@type": "Thing",
                            "name": "Ingestion Trials",
                            "description": "Neural Command's public 21-day trials measuring how ideas enter machine memory, instrumented with canary phrases."
                        }
                    ]
                }
            ]
        }
    ],
    "representations": {
        "html": "https://nrlc.ai/research/co-occurrence-case-study/",
        "markdown": "https://nrlc.ai/research/co-occurrence-case-study.md",
        "json": "https://nrlc.ai/research/co-occurrence-case-study.json"
    },
    "publisher": {
        "@id": "https://nrlc.ai/#organization",
        "name": "Neural Command, LLC"
    }
}