[b-hebrew] Sorting Hebrew

Peter Kirk peter at qaya.org
Tue Nov 28 13:04:09 EST 2006

On 28/11/2006 03:20, Bob MacDonald wrote:
> One of the technical things that makes life difficult for an English speaker
> is the finding of words in a lexicon.  First, some of them are spelt
> multiple ways - with and without mater; second, of course, the prefixes make
> parsing mandatory before figuring out which initial letter to search under.
> In this area Joel Hoffman's recent note on plurals and future and persons
> was very helpful.
> But the worst aspect of the BDB is that several words are not in what an
> English speaker would consider a natural order at all - loads of them
> beginning with Mi and variations appear in almost every section of the
> alphabet, (one gets used to a large exception) but now - for Psalm 89, the
> word TBL is listed for whatever reason under Yod!
> What is going on in the minds of those who sort the fiery letters? Any
> clues?  Of course I won't mention the ubiquitous Strong numbering system
> that reduced many English clerics to lunacy when teaching what little they
> learned in their three week mandatory Hebrew courses.
> I wrote a quick and dirty algorithm to derive a root from a word in the
> psalms. This combined with a random search mechanism helps me find what I
> have already 'done'. What a magical forest! I doubt I will ever publish a
> Hebrew sort algorithm - but someone must have this trouble today in the
> Unicode world.
> Did the ancients have a concept of sort? Or even dictionary lookup?
Good question! We do know from the acrostic psalms etc that the ancient 
Hebrews had an alphabetical order. But beyond this I think we are 
guessing, at least for Hebrew.

I have a less dirty way of finding the root of a word in the Psalms, or 
elsewhere in the Hebrew Bible: I look it up with my Bible software, 
which includes data on the root of every verb form in the Bible. Of 
course this data is not 100% because there are some ambiguities (as in 
some of the examples I just quoted for Karl), but it is more reliable 
than any purely algorithmic approach might be.

As for sorting Hebrew in Unicode, there is a standard Unicode sorting 
algorithm with a default order for Hebrew, although I think the 
secondary sort by vowel points is somewhat arbitrary. I have been 
somewhat involved in fine tuning this. But it is designed primarily for 
modern Hebrew, which means that sin and shin are sorted together at the 
first level. This was interestingly enough the sort order used by 19th 
century dictionaries including Strong's, and so Strong's numbers are 
ordered on this basis. But 20th century dictionaries of biblical Hebrew, 
from BDB onwards, have sorted sin and shin as separate letters at the 
top level. The Unicode algorithm also provides for tailored sort orders 
for specific languages or other requirements, and one could be set up 
for biblical Hebrew as in BDB. But software will continue to sort by 
default in the modern Hebrew order as for example Microsoft Excel 
already does (which was in fact helpful to me in manually finding 
mininal pairs between sin and shin to respond to Karl), so it may be 
that the 19th century order will come back into use in the 21st century.

Peter Kirk
E-mail:  peter at qaya.org
Blog:    http://speakertruth.blogspot.com/
Website: http://www.qaya.org/

More information about the b-hebrew mailing list