Tuesday, 8 December 2015

The semantic range of giving

I have a pattern of assignment of gloss that is evolving in my reading of the Hebrew Scriptures. A gloss is not a translation. Defining anything is difficult. I would rather call my 'translation' at this time 'a reading' of Scripture. It's not an opinion because I have a host of opinions and I don't necessarily want to be ruled by them. Some of my opinions are unruly. In 'reading' and reporting I want to open the faces of the texts to reflect their many facets. There are many readings. When I am finished, it will be a translation and I hope it retains the ability to be read and to reflect many faces.

The pattern is at the moment as follows:
  • A Hebrew stem can be reflected in many different English glosses.
  • An English gloss should represent only one Hebrew stem.
This is not a one-to-one relationship. Some words have one gloss throughout, but most words do not. What I am doing is striving for a one-to-many relationship. Exceptions.
  • Prepositions and conjunctions are fluid and contextual in both languages. Subtlety required.
  • Homonyms in English must allow for multiple Hebrew stems.
  • Helping verbs in English are allowed as freely required for grammar.
  • An English gloss + preposition is considered distinct from the English gloss alone.
  • Word games may take precedence over these rules of concordance.
So what do you do when give means take, or give means put, or give means set up?

There's the rub - it lies in the little word 'means'. I do not translate for 'meaning' if that means 'demeaning' or simply selecting the middle value or being mean. The 'meaning' of a text is in the relationship set up, given, from reader to reader, from one unspeakable to another. I seek to perceive many possible 'meanings' by preserving the patterns that I learn as I read more and more.

When I'm stuck, what I do is search until I decide that an English homonym or exception will need to be defined to my algorithm to allow the exception in some cases. So give can be take as in 1 Samuel, 'Do not take your maidservant as the face of a daughter of worthlessness,'

And give can be yield, and there are several others at the moment, though how long some of the rare ones will last is moot. Sometimes I look back at the list and wonder if it should stand. But I try at all costs to avoid a many-to-many situation of source stem to gloss, because that simply loses any if not all discipline, and the reading becomes too subject to my face only.

This self-justifying thought was produced as I approach the creation of the tabernacle and the ark. (I am going to call Noah's 'ark' something else - big boat maybe.) Here I am talking about the ark that is the intimate presence of God with the people. The language is delicate and the rules are like a father showing his child how to make something. Not a bossy know-it-all father either.
וְנָתַתָּ֖ אֶל־הָאָרֹ֑ן
אֵ֚ת הָעֵדֻ֔ת אֲשֶׁ֥ר אֶתֵּ֖ן אֵלֶֽיךָ
And you will yield to the ark,
the testimony that I will yield to you.

The primary gloss for נתן is give. Is this gifting, to yield to us the 'testimony'? Is it our gift to yield the tables of the law to the box under the mercy seat?

Law is then yielded to mercy and compassion. (Now that is subversive.)

There are of course many other aspects of reading and reporting at this stage that are equally difficult.
  • The tense in which a verb is expressed is moot. Almost anything goes in poetry and also in prose. The basic rules of course are observed, but there are many exceptions. English and Hebrew each have a host of possibilities in tense, aspect, mood. It has to do with telling the story or allowing the story to have its provocative and subversive impact. Often it is a matter of stepping back from the obvious. It is a political game of love.
  • Neither language has equality of gender at its base. Genders are of course not equivalent in all respects, but legally the masculine may include the feminine when no other singular allows representation. I often use 'it' as an inclusive personal pronoun, but I am far from consistent. Grammatical gender, e.g. the feminine noun, is not a feature of English, but it is of Hebrew and the feminine particularly can represent metaphor (like the boat in Jonah who 'was reckoning she was a wreck'). The degree to which language is play must be considered. Metaphor is unavoidable of course.
When I have the computer assign glosses automatically, the algorithm takes this precedence:
  • assign the glosses from read to unread sections by phrase, beginning with 6 in a row and reducing progressively to 2 in a row. As the assignment is made, correct the stem and semantic domains in the unread section.
  • for single words in Hebrew, assign the gloss only for matching word and stem and semantic domains, preferring first from the same chapter, then from the same book, then any book with the exception of Daniel (Aramaic) and the chapters that are acrostics.
So far the algorithm stats are improving each time I run it. Right now the complete unknowns have reduced over the last 6 weeks from over 75,000 to nearly 60,000. The knowns are now over 72,000, and the knowns if I include guesses by repeated phrases are over 107,000. That leaves 65% of the Hebrew 'words' remaining.  (Hebrew words are more often than not English phrases because Hebrew is an enclitic language. I.e. its components stick together to form a lexical word).

For interest, this is the current list of phrases and their count, which arise from the word נתן in my data. It could use a few tweaks.


No comments:

Post a Comment