| Before you can tabulate, you must have at least
one entry annotated.
Tabulate. In annotating,
you express your judgments about what forms in an entry could be
considered as having come from the same protoform, by tagging them
as members of a group of potential cognates. You also specify how
the segments in the forms of each group should be lined up for
comparison, by inserting Indel and Ignore symbols as needed.
Tabulate displays the groups of
potential cognates set up on that basis, one at a time. It slices
the group by segment positions to yield correspondence sets.
For each correspondence set, you express two more judgments:
- What protosegment in the
earlier language might this correspondence set represent, at
least tentatively?
- What is your best guess about
the phonological environment in which it might have occurred in
the protolanguage?
On the basis of those
judgments, Wordcorr selects the most likely place to insert each
correspondence set into the results structure. It's not always the
best place; that is for you to decide when you get to the next
stage, refinement of the
results. Go there after you have tabulated all the groups in the
entry you're working on. |
INDEL is a term coined by David Sankoff and Joseph Kruskal
for a segment that was either INserted into or DELeted from a
string of segments. In comparing forms like bua and buwa,
did the w just appear in an earlier form bua, or did w
drop out of an earlier buwa? Until we get enough other
patterns to estabish the direction of change, we're guessing. So
"/", the Indel symbol, means "insertion or deletion,
we still don't know which."
IGNORE is symbolized by
"." It is used to represent missing or irrelevant data.
Both concepts are explained more
fully in Wordcorr Help. |
