dnevnikmelochej

• First, this task required getting a list of standardized glosses with which we would append to each of the database entries.
• So a preliminary list of standardized glosses came from a nice printed-out list from Sutton & Walsh (S&W). Meanwhile, Professor B and I discussed what categories and parts of speech there should be in the database-- this was an issue during the school year that was known for a while and needed to be addressed.
• Manual input of standardizing glosses was initially unwieldy, since the items were not initially easily accessible on the computer.  There were also many terms that was a "near synonym" with another S&W entry, and I started a spreadsheet containing such a list of near synonymous glosses. 
• FInding the digital S&W files (as spreadsheets) and doing searches on the computer made things much more manageable, but there was still room for improvement.
• We soon noticed that each entry in the S&W list could be easily tagged for semantic category and part of speech data as well, and I managed to do such tagging manually.
• Professor B decided to incorporate the S&W list (tagged by category and part of speech) directly into our database, and gloss standardization became significantly easier, since putting in any item of this S&W list will automatically fill in the category/part of speech information. I particularly liked the fill-ins, because it was very intuitive although it takes a little while to get used to. After all, you need to know what the glosses look like. There are some general patterns and principles, like "bird" versus "bird spp.", or "teeth" being glossed as the singular "tooth", plural "things" instead of "thing", but some other patterns were more arbitrary (like "light" as a noun versus "light (weight)" as an adjective). Another problem of automatic fill-ins occurred with cases like "dream" (which can either be a noun or verb) and "drown" (either transitive or intransitive) where sometimes you would prefer that they leave the part of speech information blank.
• It soon became clear that S&W would not be the final word for many entries in the database that would remain unglossed because they didn't fall into any good standardized gloss in S&W. For instance, the first version of S&W lacked in its vocabulary many function words ("because"), discourse particles ("hello", "goodbye"), abstract concepts ("god", "dreamtime"), specific terms for organisms ("oyster"), and all the complex kinship relations ("mother or mother's sister", "uncle-cousin pair"). Fortunately, the database is very good at allowing new entries to be manually added, and this even applies to the S&W list which continued to grow day by day.  It quickly became very exciting when we could surpass S&W's established standardized glosses and find some other terms worthy of standardization.

• I briefly talked to Professor B about phonemicization, and how hard it was to correct the inconsistencies in orthography of various sources during phonemicization with no knowledge of the phonemic inventory of differing languages.  It would be nice if in the database each language/dialect had a nice reference to an IPA chart with interesting features of the language noted (number of rhotics, palatal/apical distinction, etc).
• Although we have spent this summer trying to rigorously delimit the "parts of speech", "generalized glosses", "phonemicized form", and "semantic field", there are still other fields in the database that, although perhaps non-essential for the purposes what we are interested in, also needed to be delimited with the same vigour.
  •  "Variant" vs. "synonym" field.  I've put orthographic variants (i.e. "wadly" instead of "wadli") under the "variant" field, while "synonyms" are more obviously different words with an approximately similar meaning.
  • "Source page" vs. "source note" field.
  • "General note" vs. "other note" field. Under "general notes" I would include parts of the extended definition of a word-- like if one is talking about a ritual, then the general note will go into more depth about how the ritual is done, who does it, etc.  The "other note" field would be for other observations, like "reduplicated form is X, meaning Y"
  • Also "PScopy" vs. "Part of Speech".  I know that "PScopy" is in fact a copy of the "Part of Speech" field... but when should we get rid of the copy?
  • "Form note" versus "paradigm note" field.

So today began the start of my last week in New Haven.  Professor B and I had a talk this morning about what we would like to get done before the beginning of the fall semester, when other Yalies will start helping out on the grants again.

We set several things in motion, such as tracking which languages should be "adopted" by various student researchers in the fall, getting sound files appropriate for my senior thesis, and seeing what work I could do in the meantime. Today I worked on putting in some more Duungidjawu entries that Barry had started.  I think I got around 95% finished with this language at the end of the day.