standards for translation?

[Moustafa Mounir Elqabbany مصطفى مني ر القباني <elqabbany at sunnipath dot com>]

Assalamu alaikum.

This is my first post to the doc list.  If these matters have been
discussed previously, please point me to a link.

Insha' Allah, as per Ahmed Al-rasheedan's request, I'm working on the B
list.

I have some concerns about formalizing and standardizing the
translation methodology wrt the word list.  We really have to design
this properly if we want the
wordlist to be truly useful.  An ad-hoc translation methodology will
result in the dictionary not being taken seriously by academics or
professional translators.  I have actually been quite disappointed with
many dictionaries in the past.  IMO, it's better to to not produce a
dictionary than to produce a bad one.  A dictionary should be a tool
that an
educated student of the language can use to accurately understand or
translate documents.  When I look up the translations, I expect
to find consistency and academic rigour.  

Here are my thoughts.  Some of these are so obvious, you might wonder
why I'm mentioning them.  I'm doing so because it allows me to spell
out important distinctions that become not-so-obvious later on.

1. In this project, there is a source (English) and a target (Arabic) language.  The entries of the source language should be organized according to the conventions of native dictionaries in the source language.  For example, in English dictionaries (as opposed to traditional Arabic dictionaries), one typically finds an entry for many inflections of the same word in the dictionary.  This is currently being done in the wordlist project, which is a good thing, alhamdulillah.
2. A word can have more than one definition.  e.g., "to run up against an opponent", "to run around", "that was a good run".
3. The wordlist project should aspire to provide as many translated definitions as possible.
4. As a corollary to the above, it is also absolutely critical to include indications on how to properly use phrasal verbs, since in both Arabic and English, the definition of a verb can change depending on the preposition following it.  e.g., "I ran into him", "I ran up to him".
   
5. The goal of the wordlist is not to teach people Arabic morphology (صرف).
6. If an entry can act as different parts of speech, such as a verb, noun, adjective, particle, etc., then every definition of every part of speech that applies to the word should be translated (ideally) .
   
7. Verbs are trickier than nouns, because an uninflected verb often corresponds to multiple conjugations (depending on person, mood, etc.) of the verb.  Thus, when translating uninflected English verbs (e.g. "run"), we should only translate the infinitive form, which should be fully vowelized, and include both the present and past tenses.  The benefit of translating the infinitive form is that it is the most generic and one will usually be able to look up all other forms based on it.  It should be fully vowelized to avoid ambiguity.  Finally, it should be in both the present and past tense because, (1) this indicates to the reader that it is an infinitive, (2) avoids scenarios where a two verbs share a common past tense but differ in the present, and (3) clarifies difficult-to-determine past or present tenses, such as when one of the letters is an alif, waw, ya, or hamza.  I realize that some of the justifications I've mentioned in this point do seem to infringe on #5.  However, #5 was never a criterion by which I concluded that we should do anything.  Rather, it was just an added bonus.
8. The previous point does not apply to inflected verbs.
   
9. Based on the above, we must have a way to distinguish between different entries.  We should either have separate entries, properly labeled, or (better) have a convention to distinguish between the different entries.  For example, go to http://www.onelook.com and enter the word "run", then go through some of the links to online dictionaries that come up.  You will find that they distinguish between definitions that are intransitive verbs, transitive verbs, phrasal verbs, nouns, adjectives, etc.
   
10. When the inflected form of a word is a straightforward application of the definition of the word to its inflection, then this should be indicated as such.  For example, for "ran", we should indicate that it is a past tense verb (with a simple code that we list in a legend) and provide a hyperlink to the particular subentry of the uninflected word.  In some cases, the inflected form carries a meaning above and beyond a straightforward application of the definition to the inflection, such as "runner", which is a kind of shoe, in addition to being the active participle of "run".  However, the active participle of "run" only applies to certain definitions of the verb, and "runner" the shoe is not a direct application of *any* of the definitions of the verb.  The dictionary entry for "runner" should make all of this clear.
11. We can do the above by carefully thinking through the codes and conventions we can use for dictionary entries.  Hopefully, there will be a concise way to represent this without confusing the reader.  I'm not very familiar with tools like kbabel yet, so maybe they already deal with some of the issues here.
12. We need a policy for removing translations from the wordlist.  I have already come by a translation that I believe to be flawed.
    

-- 
Moustafa Mounir Elqabbany مصطفى منير القباني
(The above message uses Unicode (UTF-8) encoding.)


begin:vcard
fn;quoted-printable:Moustafa Mounir Elqabbany =D9=85=D8=B5=D8=B7=D9=81=D9=89 =D9=85=D9=86=D9=8A=
	=D8=B1 =D8=A7=D9=84=D9=82=D8=A8=D8=A7=D9=86=D9=8A
n:Elqabbany;Moustafa Mounir
email;internet:elqabbany at sunnipath dot com
tel;work:604.542.1830
tel;cell:604.710.6994
x-mozilla-html:TRUE
url:http://www.transparen.com
version:2.1
end:vcard