To read further, use the MENU button at the top of the page to access a drop-down list of the main areas of the website. Hover over Bayala Databases to access a second drop-down menu listing the other pages which describe the operation of the databases.
2. THE ‘WORD’ FIELDS
The 'Word' fields appear at the bottom of the OVERVIEW layout in Fig 2.1 immediately adjacent to the Source fields.
Fig. 3.5 Australian, Respelt and ‘No Hyphens’, with text mark, transcription indicators and a ‘prefix’ field
The Word fields consist of the following, treated below briefly in turn:
AUSTRALIAN (grey)
The ‘Australian’ field presents the original language record. It is shown in bold Times Roman typeface. (If it is not bold, that is a chance error of no consequence.)
RESPELT (light brown)
The 'respelt' field is partly obscured in Fig.3.5, but clicking in it will reveal its content:
Fig. 3.6 Clicking on the obscured ReS field reveals its content
The original purpose of the respelt field (‘rsplt’ or ‘res‘) was to respell words in a consistent style following conventions used by linguists, while retaining word division using hyphens, and including the use of double letters, whether vowels or consonants, and occasional accents (diacritics), if they occurred in the original and appeared useful. It was later decided to use the respelt (res) field differently.
​
The respelt field had been used for respelling with hyphens, double letters and occasional accents, and the 'No Hyphens' (NoH) field for respelling without hyphens or any such extras. The new role was for the respelt field to be a simplest respelling of the whole word in the NoH style, while NoH would now show the stem of a word with suffixes placed in subsidiary analysis fields. This change in purpose became necessary because otherwise there would be no way to search the database for, say, badabami, the word having been separated into its constituent parts using the NoH and analysis fields. (Before making this change, what would have appeared in the respelt field would have been bada-ba-mi.)
Fig. 3.7 Respelt field showing the full record while the NoH field shows only the stem, with constituent parts in analysis fields
The pair of examples in Fig. 3.7 show the future tense marker ba in the dark blue ‘TNS’ field, and the nominative pronoun mi (thou, or you-singular) in the final, purple, field, which is sometimes indicated as 2sgNOM in the ‘Category’ fields in the database.
NO HYPHENS (brown)
The ‘No Hyphens’ field (or ‘NoH’), while often the same as the respelt field, may show the stem or root of a word, with other parts split off into analysis fields:
Fig. 3.8 Difference between respelt and No Hyphens fields
The respelt and No Hyphens fields aim at the simplest possible presentation of a respelt word, including avoiding the use of double letters and hyphens, to maximise the number of ‘matches’ when a search is undertaken. These fields are not intended to indicate a ‘correct’ way to respell indigenous words.
One benefit of applying the simplest respelling consistently across all records in the Bayala databases is that however words were spelt by the original recorder word matches can be identified across languages from the Tiwi Islands in the Northern Territory to Tasmania.
RESPELLING PRINCIPLES
The respellings adopted in both the res and NoH fields follow some basic principles. These include:
​
—only three vowels, /a/, /i/ and /u/ with sounds as in Australian-English ‘path’, ‘suite’ and ‘put’.
‘e’ and ‘o’ are never used in the respellings in these databases
—/g/, /d/ and /b/: voiced consonants used exclusively in favour of the unvoiced equivalents ‘c’/’k’, ‘t’ and ‘p’. Aboriginal languages lack the voiced and unvoiced contrast of English and many European languages
—dy used for the sound of ‘j’ (and of ‘g’ as in ‘ginger’)
—never have two vowels together: whether to represent a long sound (aa, as for the a in ‘far’); or as a diphthong, as the sound in the English word ‘eye’ and first-person subject pronoun ‘I’, spelling this /ayi/ and not /ai/
Note: The recording of Australian indigenous languages was not assisted by the particular language in which it was mainly done, English, with its less-than-uniform spelling characteristics. Had the recording language been one such as Italian, where every letter or combination of letters may be interpreted in one way only, there would have been fewer difficulties in understanding the sounds the original recorders intended to portray. The examples given in Fig. 12.5 showing the many spellings of the Sydney language word for ‘good’ illustrate this variability. Had the language of the multiple recorders been Italian, it is likely most would have rendered the word in one way only—bugiari—to be transcribed in these databases as budyari.
​
PHONETICS
There is no column that attempts to present a truly phonetically accurate rendering of any word, although such a column could be added. It might be of interest but would not help—for the more specific and individualised the transcription, the fewer ‘matches’ will occur, and word matching is one of the principal valuable features of these databases.
LOCAL RESPELLING SYSTEMS
There is also no column that reflects the spelling choices of any local indigenous group. Some such groups prefer the adopting of ‘k’, ‘t’ and ‘p’ in place of /g/, /d/ and /b/. To adopt such respelling would also reduce word matches.
Some systems reproduce ‘long’ vowels by doubling them. It is known (from William Dawes’ observations) that in the Sydney language ‘to see’ had such a long vowel, rendered by Dawes as ‘naa’. However, consistent with the ‘no doubles’ principle, in this database system only na is used. Part of the reason for this avoidance of portraying long vowels by doubling is that while ‘naa’ might be known, few other Sydney language words featuring a long vowel are known with certainty, and a ‘no doubling’ principle avoids guesswork that would otherwise be necessary.
Australian indigenous languages have a way of pronouncing ‘n’ and ‘d’ that is different from English and sometimes rendered ‘nh’ and ‘dh’ by transcribers. For such occurrences, these databases may use a capital N and D. These are strong, obvious, visual clues to the corresponding sounds, and harmlessly employed. This is because the computer program, Filemaker Pro, does not distinguish between lower case and capital letters and as a consequence this device has no negative impact on the sorting strength whatsoever.​​
Fig. 3.9 Capital letters used in the database to distinguish long vowels or special consonants
Another device used, taking advantage of Filemaker Pro's ‘blindness’ to the difference between capital and lower case letters, is in the way the /ng/ sound occurring twice in ‘singing’ is presented. This is a common sound in Australian indigenous languages as in English, but, unlike in English, the sound can often begin a word in Aboriginal languages. The way of showing this sound in phonetics is to use the symbol Å‹. However, sometimes the ‘ng’ combination does not sound that way, both in English and in Australian languages: compare English ‘finger’ with ‘singer’. The same applies in Aboriginal languages, with the sound portrayed by ‘ng’ sounding as Å‹ and sometimes as n-g. In the Bayala databases when the sound Å‹ is intended, the ‘ng’ is presented in bold type as in manyimangadyiminga in Fig. 3.8. On such occasions when the ‘g’ is separately pronounced, the device used is to write -ngG-, or -nG-, with a capital G to indicate the separate pronouncing of the letter, as in the following:
Fig. 3.10 Use of ng and G to show pronunciation indicated by spelling of original record
When there is no indication from the original record whether the sound Å‹ or -n-g- was intended, a word (such as wanga-wanga, a type of pigeon) might be written as wanga wanga, wanGa wanGa or wangGa wangGa, the first option without either bold type or capitals indicating that it might be pronounced either way. See further 11. Bayala Database Conventions.
TRANSCRIPTION INDICATORS (black and dark grey)
The ‘Transcription indicator' fields (Tr1, Tr2), black and dark grey respectively, provide keys to the particular transcription rule followed for the example concerned.
Fig. 3.11 Transcription indicators, and the explanation of the first example displayed in the black/grey transcription bars above
About a hundred rules are used in all, whereas for an Italian-based spelling system it might have been that none may have been needed. Transcription indicators and rules are dealt with in 7. Supplementary Information and 12.3 Related Databases.