Stop transcriptions of the Sydney Language (Gombeeree) in contact-era vocabularies, 1790–1902: four-recorder comparison set and compiled corpus

Stop-letter data for a study of the reliability of contact-era orthography in the Sydney Language (community name Gombeeree), the language of the Port Jackson district, New South Wales. Part A, the four-recorder comparison set, aligns every lexeme recorded by two or more of the four largest independent vocabularies written between 1790 and 1796 (Collins, Dawes, Southwell and Paine) stop by stop: 119 positions in 85 glosses, with each recorder's form as written, source notes, place, position, environment and agreement codes. Part B, the compiled corpus, contains 615 stop tokens from documentary sources of 1790–1902, taken from the first five sequential categories of the wordlist compiled by Troy (1993) and traced to the spellings of its designated sources, coded for letter, voicing as written, position, cluster and geminate status. The analysis workbook reproduces every count, agreement coefficient (Krippendorff's alpha; Cohen's kappa with bootstrap intervals), model fit and table from these data with live formulas. Letters are coded as written; no spelling is normalised. The README gives file contents, coding rules and the mapping of source labels.