[Tshwanelex-l] Fwd: showing genders / past tenses

Claire Bowern claire.bowern at yale.edu
Sun Jun 12 12:42:06 EDT 2011

---------- Forwarded message ----------
From: David Joffe <david.joffe at tshwanedje.com>
Date: Thu, Mar 31, 2011 at 3:29 PM
Subject: Re: [Tshwanelex-l] showing genders / past tenses
To: tshwanelex-l at mailman.yale.edu

> The dictionary I am preparing needs to show, not only genders in
> Norwegian, but all allowable forms of each word. I want the forms to
> be searchable by the end-users (not just the lemmata). This applies
> for past tenses and all other parts of speech.
> I am assuming I need to add new attributes, right? Could not follow
> the Users Guide on the way to add attributes. Sorry!

This part is a bit non-trivial, and there can be more than one
'right' way to solve a particular problem. Some rules of thumb:

- If there can be more than one of some type of info, that usually
suggests you want to first add an element, and add your new
attribute to that element. This is because for each attribute of an
individual element in the Tree View, there is one (and only one)
attribute value for that element::attribute. If there is only one of
some type of information (for a given 'owner' element), then that
information can be an attribute.

For example if each headword has only *one* Norwegian 'gender'
associated with it, then the simplest is to create a new attribute
on the 'Lemma' element, e.g.: Lemma::Gender

If there can be more than one 'allowable forms' for each headword,
then you may want to first create a new container element for
holding one 'allowable form', e.g. "AllowableForm". Then, allow the
AllowableForm to be a child of a Lemma element by using the DTD
editor to add a 'zero or more' 'child relation'. Then, add an
attribute to the AllowableForm element to contain the actual *data*
for each allowable form. It can have the same name, e.g. also
"AllowableForm", so the full element::attribute 'notation' would be

Then in the Tree View you will be able to create structures like:

Lemma: LemmaSign=judgment, NorwegianGender=m
  +--- AllowableForm: AllowableForm=judgement
  +--- AllowableForm: AllowableForm=jujmint

So, elements form the 'skeletal structure' of the Tree View
hierarchy and can have 'zero or more', while attributes fill in the
'content', and there is one of a particular element's attribute for
each element, and one only (though it may be left empty, generally).

Disclaimer: The above may or may not be a valid or good solution,
depending on what exactly you mean linguistically by 'allowable

That's just a very quick 'beginners primer', not sure if that helps.

Of course, if there were only two possible 'allowable forms', one
could also have just created two AllowableForm attributes on the
Lemma element, e.g.:


That's a simpler solution, especially for beginners, but is
generally a slightly inferior solution to using elements properly.

 - David

Tshwanelex-l mailing list
Tshwanelex-l at mailman.yale.edu


Claire Bowern
Associate Professor
Department of Linguistics
Yale University
370 Temple St
New Haven, CT 06511
North American Dialects survey:

More information about the Tshwanelex-l mailing list