2.1 Previous research on dialect boundaries
2.1.1 Research by D. Lithgow
Dr. David Lithgow and Philip Staalsen of SIL conducted a language survey of the D’Entrecasteaux Islands
9
of Milne Bay in 1964. They elicited wordlists from the survey area and then used these wordlists to produce an initial classification of the languages in the area, along with dialect divisions. They
considered two speech communities to be “separate languages and mutually unintelligible if the percentage of cognates between lists of vocabulary items is less than 81” Lithgow and Staalsen
1965:4. They joined dialects together as members of a single dialect group if the similarity of these dialects to each other was greater than their individual similarities to the central dialect.
10
The dialects of the Galeya
11
language were thus grouped into two dialect groups: Northern Gameta, Ulua, and Basima and Southern Sebutuya and Garea, with cognate percentages shown in table 3. See Appendix
C.1 for further details on wordlists collected by Lithgow. Table 3. Cognate percentages—Lithgow and Staalsen
a
Galeya
92
Sebutuya
82 82
Basima
79 77
90
Ulua
75 73
87 94
Gameta
a
Based on research from 1964; this table appears in Lithgow 1976:459.
Lithgow also investigated the phonological and grammatical structure of the languages of the D’Entrecasteaux Islands, summarized in his article “Austronesian Languages: Milne Bay Province”
1976. He does not, however, make distinctions between dialects; he simply refers to the Galeya language as a whole.
2.1.2 Research by Troolin and Engkvist
As mentioned in section 1.3.1, David Troolin, SIL surveyor, and Leif Engkvist, SIL linguist working in the Molima language on Fergusson Island, completed a preliminary survey and report of the Galeya
language in October 1996 while investigating the dialects of the Molima language. They spent three nights in the Galeya area and visited five villages along the coast. In the report they also suggested that
there are two Galeya dialects, Galeya and Basima. In the Basima dialect they included Gameta, Aomotana, Ulua Wagili’u, Lauoya, Basima Gubayawe, Eiatoto-unu, Duduna, Yaya Point, and possibly
Wadalei and Bowada.
12
They included in the Galeya dialect Sebutuya Watuyuwa and, possibly, Didikaia and Keboluai.
13
9
Goodenough, Fergusson, and Normanby Islands.
10
The central dialect is defined as the dialect having the most shared cognates with all other dialects of the language.
11
Galeya is referred to as Garea by Lithgow and Staalsen in 1965 and Galeya by Lithgow in 1976.
12
Bowada was not listed by respondents in the 2004 survey. On old maps it is located on the coast, north of Sebutuya Bay. Bowada was mentioned by the pastor in Sebutuya who said he had been there for training.
13
Neither Didiakaia nor Keboluai were listed by respondents in the 2004 survey. On old maps Keboluai is located inland from Sebutuya Bay, south of the current location of Bibio. Similarly, old maps locate Didikaia as being
somewhere on the border with the current Momo’awa and Sebutuya wards, or perhaps in Momo’awa ward.
Table 4. Galeya villages by dialect—Engkvist and Troolin 2002
Basima dialect Galeya dialect
Gameta Galeya
Aomotana SubutuiaWatuyuwa
Urua UluaWagiliu Didikaia
Lau’oya Keboluai
BasimaGubayawe Eiatoto-unu
Duduna Yaya Point
Bowada In addition to the above villages, the report listed more than seventy hamlets under five wards:
Lauoya, Basima, Duduna, Sebutuya, and Momo’awa. Engkvist and Troolin reported that the Basima dialect villages are all within a few kilometres from the coast, although Momo’awa ward has some inland
villages. Unlike Lithgow, whose dialect division was based purely on lexicostatistics, Engkvist and Troolin
based their dialect conclusions on both lexical and sociolinguistic data. They compared the wordlists and identified two dialects Basima and Galeya with an average of 75 percent lexical similarity between
them.
14
Table 5. Lexical similarity percentages—Engkvist and Troolin 2002
Yaya
95
Gubaiyawe Basima
90 88
Wagiliu Urua
78 73
75
Gameta
Engkvist and Troolin’s percentages were significantly lower than Lithgow’s e.g. they gave a percentage of 72 percent similiarity between Basima Gubaiyawe and Sebutuya compared to Lithgow’s
82 percent.
15
However, they came to the same conclusion, that there are two dialects, Basima and Galeya. In addition to the linguistic data, they also considered sociolinguistic factors. They reported that
the people in the Basima-speaking villages said they could understand and communicate with the Galeya people. This perception of being one group affected mutual intelligibility, and placed them in the same
language, although different dialects.
In addition to their conclusion on dialect boundaries, Engkvist and Troolin also charted characteristics of the language in terms of grammar and phonology, though once again, as in Lithgow’s
work, this was a general overview of the language as a whole with no reference to possible dialect differences.
14
Engkvist and Troolin used a rule of 50 phonetic similarity in order to calculate the lexical similarity percentages. “Phonetic similarity is defined as sounds that display a ‘reasonable amount’ of physical similarity Crystal
1987:161, which in this report, was based on the IPA character chart. If two sounds were more than two squares away, either vertically or horizontally, they were classified as different” Engkvist and Troolin 2002:19.
15
See Engkvist and Troolin 2002:19 for a possible explanation of discrepancies.
2.2 Reported dialect boundaries