BeforeUp

Here is an example of an entry from a source (titled Lenje Handbook) that we have acquired:

Let’s see what you, as a PanLex developer, might typically do to this entry. (Don’t worry if it’s partly unclear now; explanations are on other pages.)

Step 1

First, you select the PanLex-relevant information, standardize it, and put it into the form of a row in a table (tabularization). Here are some possible ways to do this:

Simple:

chona    mat=bed

A little more detailed:

chona    n    mat=bed

Even more:

chona    n    mat=bed    place to sleep in (mat, bed, etc.)

Really ambitious:

chona    n    mat=bed    art-314:location_of:eng-000:sleep

Super-ambitious: combine the last two above into a row with 5 columns.

Step 2

Second, you convert the table into a sequential format (serialization). If you chose the simple version above, the result looks like this:

mn
  dn
    leh-000
    chona
  dn
    eng-000
    mat
  dn
    eng-000
    bed

If you were really ambitious, the result looks like this instead:

mn
  dn
    leh-000
    chona
    dcs2
      art-303
      PartOfSpeechProperty
      art-303
      CommonNoun
  dn
    eng-000
    mat
  dn
    eng-000
    bed
  mcs2
    art-314
    location_of
    eng-000
    sleep

Step 3

Third, you submit your serial file for importation into the PanLex database.

After

After you have performed steps 1, 2, and 3, there are new records in several tables of the database. If you chose the super-ambitious version, here is what you have added from this entry:

• If “chona” in language variety leh-000 (Lenje) isn’t already an expression in the database, a new record for that expression in table ex.
• A record for the new meaning of source leh-eng:Madan in table mn.
• 3 new records for denotations of that meaning, assigning it to each of the 3 expressions (“chona”, “mat”, and “bed”) in table dn.
• A record for a definition of that meaning in language variety eng-000 (English) in table df.
• For the “chona” denotation, a record in table dcs classifying the denotation as belonging to class “CommonNoun” in superclass “PartOfSpeechProperty”.
• A record in table mcs classifying the new meaning as belonging to class “sleep” in superclass “location_of”.

As a result, users of PanLex can find new information. For example, a user wanting to translate “အိပ်ရာ” from Burmese into Lenje can get “chona” as a distance-2 translation.

What did you do?

To understand fully the example above, you need to study the documentation in this assimilation section. But here are a few points:

• You decapitalized “Chona”, because capitalizing the Lenje expressions is part of this source’s style, not part of any Lenje orthographic standard.
• The form “shona”, the source says in its introduction, is a plural form, so you omitted it, knowing that PanLex is a database of lemmatic translations.
• You also omitted the words introduced by “Cf.”, as being of little value to PanLex.
• You chose whether to omit “a place to sleep in”, treat it as a definition, treat it as a meaning classification (“really ambitious”), or both (“super-ambitious”). The meaning classification takes more time to create than the definition, but also has more potential value.
• Are “bed” and “mat” good translations of “chona”, or only examples? You could have refused to make denotations from them, but this could in practice prevent users from finding a Lenje translation of “bed”. By not demanding “perfection”, you brought the world one step closer to being able to express any concept usefully in any language.