CLDF dataset with data and supplements for Barlow “Loss of colexification of ‘hand’ and ‘five’ in Austronesian languages”
If you use these data please cite this dataset using the DOI of the particular released version you were using
This dataset is licensed under a CC-BY-4.0 license
Words for the concepts 'five' and 'hand' in Austronesian languages have been collected from four datasets, described in the ContributionTable. Since some entries in these datasets contain multiple forms for a single concept, and since forms were aggregated at the language level (with forms for dialects taken as forms for the parent language) and across datasets, often more than one form per concept in a given language was attested. If multiple forms were attested, one was chosen so as to maximize the potential for finding colexification. In other words, the pair of forms selected for a language is the one closest to exhibiting full colexification (or, failing that, partial colexification). The decision was made so as to minimize “false negatives” (i.e., cases where there could appear to be no colexification of the two concepts, but only because there are, e.g., two synonyms for ‘hand’ in a given language and the particular dataset chose the “wrong” one).
Based on the words for ‘five’ and ‘hand’ collected in the FormTable, six features have been coded, with values reported in the ValueTable. A seventh feature details the types of numeral systems found in the languages in this dataset; it derives from Barlow (2023), updated to reflect the changes in classifications between Glottolog versions 4.6 and 5.0. The distribution of values for these features can be investigated using geographical maps.
Replacement events (i.e., rows in the ReplacementsTable) represent a probable loss of the
inherited form for ‘hand’ or ‘five’, whether in the individual history of a single language or in a protolanguage ancestral
to multiple languages, with Glottolog languoids (i.e., language subgroups or individual languages in the Glottolog 5.0
classification of the Austronesian family) serving as proxies. While the replacements table lists the name and Glottocode of this
languoid, the individual languages in our sample that fall within this designation are linked via the Glottocodes in the
Language_IDs column.
Looking up related data from different tables of the dataset is best done by exploiting the fact that any CLDF dataset can be converted to a SQLite database. The schema of this database here is described below. If, for example, we wanted to see whether the language Lenkau appears in any replacement events, we could run the following query:
sqlite> select distinct r.subgroup from languagetable as l, "replacements.csv_languagetable" as rl, "replacements.csv" as r where l.cldf_id == rl.languagetable_cldf_id and rl."replacemen
ts.csv_cldf_id" = r.cldf_id and l.cldf_name = 'Lenkau';
South-East Admiraltyand if we wanted to see which other languages are subsumed under “South-East Admiralty”, we could run
sqlite> select distinct l.cldf_name from languagetable as l, "replacements.csv_languagetable" as rl, "replacements.csv" as r where l.cldf_id == rl.languagetable_cldf_id and rl."replacements.csv_cldf_id" = r.cldf_id and r.subgroup = 'South-East Admiralty';
Lenkau
Nauna
Penchal
Lou
Paluai
As explained in the cldf/README, replacement events can be reconstructed using a more conservative or
a more liberal approach. As an example of a discrepancy between the two approaches, consider the replacement of
*qalima ‘hand’ for Bugawac and
Kela (Papua New Guinea). In this case, both Bugawac and Kela
exhibit the replacement of *qalima ‘hand’; however, their sister language (in the Glottolog 5.0 classification)
Yabem does not, so such a replacement cannot be reconstructed
to the immediate ancestor of the three languages, North Huon Gulf linkage.
However, even though they are not subgrouped together within North Huon Gulf linkage (which has a flat—i.e.,
ternary-branching—structure in Glottolog 5.0) excluding Yabem, the two languages may nevertheless have shared
in the replacement of *qalima ‘hand’. This view is reflected in the replacements table by assigning the same Replacement_Group
value hand-43 to the two (conservative) replacement events:
$ csvgrep -c Replacement_Group -m"hand-43" cldf/replacements.csv | csvcut -c Subgroup,Comment
Subgroup,Comment
Bugawac,possibly shared change between Bugawac/Kela
Kela (Papua New Guinea),possibly shared change between Bugawac/KelaThe following CLDF datasets are available in cldf: