Lexicographic timeline

Data-driven diachronic spine for the CDSL inventory: when each dictionary was published (or spanned multi-volume years), which language family it belongs to, how large its sanhw1 lemma set is, and which titles Cologne marks deprecated. Built from committed inventory rows — not hard-coded dates in page prose.

Companion pages: Dictionary genealogy (lemma counts and inheritance — do not re-derive here), All-dictionary coverage (structure-fit scatter), agenda item V2 in ATLAS_RESEARCH_AGENDA.md.

Trust Block

Evidence: data/dictionary_inventory.csv (synced into src/data/lexicographic-structure/dictionary_inventory.csv); inventory rows, with parseable start_year (AMAR and other pre-modern markers without year columns are excluded from plots, not invented).
Limitations: orientation dates, not exact usage dating of Sanskrit. start_year/end_year are publication spans for printed editions; multi-volume works use first→last volume years. end_year=ongoing (e.g. PD) is clipped to the atlas build year for display only. Lemma sizes are sanhw1 counts, not print page counts.
Validation: checked by npm run build; inventory columns required: code,year,start_year,end_year,family,deprecated,n_volumes,sanhw1_lemmas,full_name,language_pair,author_or_compiler.
Owner repo: csl-atlas.
Next use: orient a temporal claim (PH3 / PH7 / PH8) on this spine, then verify in dictionary chapters or source-linked records. Full Heaps/era V4 ribbons need union TSV excerpts — not this page.

1. Lifespan bars by dictionary (family colour)

Each bar runs start_year → end_year. Single-year editions are thin marks. Faded bars = Cologne deprecated=yes (still historical evidence; not the preferred lookup surface). Hover for full name, language pair, and author.

Reading note. Clusters on the y-axis follow family order (earliest family first), then start year inside the family — the same “family lane” story as the agenda V2 mockup, with readable dictionary codes.

2. Family lanes (aggregated spine)

Same spans, y = language family. Overlapping bars on a lane are expected (siblings published in the same century). Useful for “which tradition is active when?” without scanning every code.

3. Volume width and multi-volume span

Bar length is still the publication span; bar height scales with parsed n_volumes (multi-volume Petersburg / encyclopedic projects stand out). Single-volume dictionaries stay thin.

Multi-volume titles (PWG, PW, PD, …) fill more of the row height; single-volume titles stay thin.

4. Lemma size at publication year

Dot at each dictionary’s start year; area ∝ sanhw1 lemmas (0-lemma reverse-direction / incomplete titles sit on the axis). This is inventory size at digitisation, not growth of Sanskrit itself. Lemma bars by code live on lexicography — this panel only places size in time.

5. Coverage ribbon — cumulative tradition steps

Thin cumulative strip: for each dictionary’s start_year (sorted), how many inventory titles have appeared by that year. Not the full Heaps/era V4 programme (that needs union headword excerpts). Use this as “when does the printed tradition thicken?”; for structure-fit vs size see dictionary-coverage.

Inventory excerpt

All dated rows used above (downloadable). Undated inventory rows (e.g. AMAR without year fields) are listed at the bottom of the table filter if present in the raw file.

Narrative version (Mermaid)

The static story arc from indigenous kośa tradition through colonial English, Bengali Skt–Skt, German philology, English consolidation, specialists, and the digital era. Data plots above are authoritative for CDSL inventory years; this block is the teaching narrative (includes pre-modern markers and CDSL tooling milestones that are not inventory rows).

For the fully-detailed version with all 50+ dates, sources, and CDSL repo links: timeline-en.md.

Source inventory: dictionary_inventory.csv. Narrative sources: DICT_PROFILE Historical background and Lineage section. Agenda: V2 data-driven timeline. CC-BY-SA-4.0.