Sensium’s Band G stack already published fifteen grape confusion cuts (Study 1) and six catalog censuses (Studies 3–8 field guide). Study 4 counted how large the process glossary is. This page is the next non-separator asset: process confusion pairs — every process confusion forks fork the glossary already encodes. It answers a different question than “which grapes twin?” It answers: which winemaking processes does the catalog tell you not to confuse on the nose and palate?
Headline finding (snapshot 2026-07-09): Across 44 processes, the glossary ships 87 confusion edges (avg ~1.98 per process). Categories: maturation 11 · special method 11 · fermentation 8 · sweetness/concentration 5 · maceration 5 · structural 4. The densest teaching families below are cream/texture (MLF · lees · oak), sparkling methods, maceration/fruit-lift, and sweetness/oxidative/flor forks.
This is still catalog research, not population miss-rates. Study 2 (topic `137`) remains volume-gated. Do not cite these forks as “which processes candidates miss most.” Cite them as the process stop rules behind Glossary and the method blogs.
Companions: Study 4 vocabulary · MLF · lees · traditional vs tank · carbonic · botrytis.
Methodology (read this before citing)
| Field | Value |
|---|---|
| Source | the process glossary — every process confusion forks string |
| Parse | Head (before `(` or em dash) + reason clause |
| Also shipped | Each process’s three `blindCues` (on the spoke export) |
| Snapshot date | 2026-07-09 |
| Snapshot | Follow the method on this page. Counts move when the catalog is edited. |
| Not claimed | Live exam miss-rates, or that every fork is a grape confusionPair |
| Study 1 | Study 4 | Study 9 (this page) | |
|---|---|---|---|
| Object | Grape↔grape separator strings | Vocabulary size (44 processes, etc.) | Process↔process (and process↔style) forks |
| Question | Which grape twins? | How big is the language? | Which method mistakes does the glossary forbid? |
Why process forks after the census stack?
Google’s 2026 non-commodity bar still applies: dated snapshot, clear “not telemetry” label. Candidates fail when they smell butter and write “oak,” or smell banana-candy and write “Pinot,” or smell honey-saffron and write “appassimento.” Study 1’s grape pairs do not own those forks. Study 4’s census says the glossary has 44 terms — it does not publish the 87 stop rules between them.
Practically: open Glossary, read process confusion forks before the grape call, and force one process fork on every flight that smells like winemaking.
Family A — Cream / texture (11 edges)
MLF, lees, bâtonnage, and oak all add weight. The catalog’s forks name the wrong cream.
| Process | Do not confuse with | Catalog reason (trimmed) |
|---|---|---|
| Malolactic conversion | Lees contact | Creaminess with biscuit/brioche, not butter |
| Malolactic conversion | Heavy oak | Vanilla/toast rather than dairy |
| Lees contact and autolysis | MLF | Creaminess with butter, not bread |
| Lees contact and autolysis | Oak | Toast, not bread/yeast |
| Lees stirring (bâtonnage) | MLF | Creamy with butter, not bread |
| Lees stirring (bâtonnage) | Heavy oak | Weight with vanilla/toast |
| Oak maturation | Botrytis | Sweetness perception with apricot/honey |
| Oak maturation | Lees stirring | Creaminess without sweet spice |
| American oak | Heavy toast French oak | Toast/smoke without coconut/dill |
| Lees aging without oak | Oak-aged whites | Similar weight with vanilla/toast |
| Lees aging without oak | Concrete-aged Chardonnay | Similar texture; can mimic neutral oak |
Drill: MLF tasting · lees · oak age · American vs French oak.
Blind cue habit: write butter vs brioche vs vanilla before “Chardonnay.”
Family B — Sparkling methods (14 edges)
| Process | Do not confuse with | Catalog reason (trimmed) |
|---|---|---|
| Traditional-method sparkling | Tank method | Softer, fruitier, less autolytic |
| Traditional-method sparkling | Ancestral method | Cloudy, less precise, often RS |
| Tank method (Charmat) | Traditional method | More toast/brioche, finer mousse |
| Tank method (Charmat) | Ancestral method | Cloudy, often with sediment |
| Transfer method | Traditional method | Finer, more persistent mousse |
| Transfer method | Tank method | Less autolytic complexity |
| Ancestral method (Pet-Nat) | Tank method | Clearer, more polished |
| Ancestral method (Pet-Nat) | Lambic / wild styles | Different category |
| Tirage | Charmat | CO₂ in tank; no in-bottle autolysis |
| Tirage | Ancestral method | No liqueur de tirage |
| Disgorgement | Tirage | End of in-bottle path, not the start |
| Disgorgement | Dosage | Sweetening step after disgorgement |
| Dosage | Disgorgement | Mechanical step that precedes dosage |
| Dosage | Naturally dry complete ferment | No liqueur d'expédition |
Drill: traditional vs tank · Champagne vs Prosecco · dosage.
Family C — Maceration / fruit-lift (11 edges)
| Process | Do not confuse with | Catalog reason (trimmed) |
|---|---|---|
| Carbonic maceration | Cool-climate Pinot Noir | Light/juicy but lacks banana lift |
| Carbonic maceration | Whole-cluster Syrah | Stemmy pepper, not candy fruit |
| Whole-cluster fermentation | Carbonic maceration | Banana-candy lift — different |
| Whole-cluster fermentation | Pyrazine-driven herbals (Cab Franc) | Herbal from grape, not stems |
| Whole-berry fermentation | Whole-cluster | Intact stems → vegetal/peppery |
| Whole-berry fermentation | Carbonic maceration | Bubblegum/banana esters dominate |
| Extended maceration | Heavy oak | Structure with vanilla/toast |
| Extended maceration | Cold soak | Color/aromatics with less tannin |
| Cold soak | Carbonic maceration | Bright fruit with banana lift |
| Co-fermentation | Whole-cluster | Floral lift with stemmy edge |
| Co-fermentation | Post-ferment blending | Less integrated |
Drill: carbonic tasting.
Blind cue habit: banana-candy → carbonic card before “Gamay village.”
Family D — Sweetness / oxidative / flor (18 edges)
| Process | Do not confuse with | Catalog reason (trimmed) |
|---|---|---|
| Botrytis (noble rot) | Late-harvest non-botrytis | Sweet without saffron/apricot |
| Botrytis | Appassimento | Raisin/fig, not honey |
| Appassimento | Botrytis | Honey/saffron, not raisin/fig |
| Appassimento | Fortification | High alcohol with spirit warmth |
| Fortification | High-alcohol dry table wines | No RS / candied-fruit lift |
| Fortification | Appassimento | Concentrated but not fortified |
| Oxidative aging | Botrytis | Dried fruit with honey/saffron |
| Oxidative aging | Flor aging | Nutty with chamomile/salinity |
| Flor aging | Oxidative aging | Nutty with bruised apple |
| Flor aging | Lees aging | Bread-y without saline edge |
| Eiswein / icewine | Botrytis sweet wines | Honey-saffron-marmalade complexity |
| Eiswein / icewine | Late-harvest non-frozen | Broader, less-precise fruit |
| Cryo-extraction | Eiswein / icewine | Same principle; natural on-vine freeze |
| Cryo-extraction | Botrytis sweet wines | Fungal dehydration + shrivel |
| Surmaturation | Botrytis sweet wines | Marmalade-honey-saffron |
| Surmaturation | Eiswein / icewine | Razor acid, lower alcohol |
| Solera | Single-vintage long aging | Tertiary complexity, vintage-specific |
| Solera | Oxidative aging | System vs regime (often co-applied) |
Drill: botrytis · Port styles.
Family E — Vessel / blend / other (33 edges)
The residual third covers amphora↔concrete egg, ripasso↔Amarone, saignée↔direct-press, field blend↔assemblage, chaptalization↔dosage, VA↔Brett, and related forks. The method table covers the full 33 — the teaching point is that process identity traps are not only MLF and Champagne. For the full pink-path drill (and the beyond-Level-3 atlas), open rosé winemaking methods · rosé beyond Level 3.
Open Glossary and walk any `special_method` or `structural` term’s process confusion forks list before inventing a grape.
How to cite this study
- Name it: “Sensium Study 9 (process confusion pairs), export 2026-07-09.”
- Link this URL.
- Name the family (cream/texture, sparkling, maceration, sweet/oxidative, vessel/blend).
- Keep the label: process-glossary forks — not grape miss-rates.
- Keep the snapshot date from the method table — do not invent a later count.
How this sits beside Studies 1–8 and Study 2
| Study 1 | Studies 3–8 | Study 9 | Study 2 | |
|---|---|---|---|---|
| Object | Grape confusion cuts | Catalog censuses | Process confusion forks | Wrong answers |
| Status | Complete | Complete + field guide | This snapshot | Volume-gated (`137`) |
After the Studies 3–8 field guide, inventing another grape aroma cut or a redundant size census would still be dishonest. Process forks are a new catalog object the glossary already owned.
A two-week drill
| Week | Focus | Habit |
|---|---|---|
| 1 | Cream + sparkling | MLF vs lees vs oak sheet; traditional vs tank vs ancestral sheet |
| 2 | Maceration + sweet | Carbonic vs stems vs Pinot; botrytis vs appassimento vs fortification |
Service: home flights · Blind.
Frequently asked questions
Is this the same as Study 1 grape pairs?
No. Study 1 edges are grape↔grape. Study 9 edges are process↔process (or process↔style), from the glossary.
Is this the same as Study 4?
No. Study 4 counts vocabulary size. Study 9 publishes the confusion forks between process terms.
Why free-text process confusion forks instead of process ids?
That is how the bundled glossary stores teaching forks today — full stop-rule sentences. Tables extract a head + reason; the full string remains the citation unit.
When does Study 2 ship?
When anonymized Train/Blind volume clears a documented threshold. Until then, prefer Studies 1 and 3–9 — and do not invent miss-rate tables.
Bookmark this page beside the Studies 3–8 field guide, open Glossary and Glossary · traditional method this week, and force one process fork before any grape poetry. Method first — then grape — then place.