Biblical Hebrew Forum

[James Christian <jc.bhebrew AT googlemail.com>]

Erm,

slow down Yitzhak. I was in Skomeroxe over the weekend. Skomeroxe is a
village with a populace of about 150 farmers in the countryside in Ukraine.
There are plenty of cows grazing in the fields, pigs snorting in their barns
and chicken running around eating everything in sight but no libraries of
pre-exilic Hebrew inscriptions. I could have asked the vodka filled and
extremely drunk man whose turn it was to pasture the village's cows but even
if the word 'library' was part of his vocabulary I doubt he would have been
able to point me to the local archives of pre-exilic Hebrew inscriptions and
would rather have poured me out 50 gram of samagonka (home made vodka) and
not let me go until I was as drunk as he was.

Now I didn't mean to offend you with my suggestion that the lacking Y in
many inscriptions could be explained by laziness. It was just a suggestion.
I understand this is your personal research and as such you may have strong
feelings about it. I'll read the article for you if you find it too
difficult to expand on why you feel that some inscriptions are not official.

Finally, you asked about measures. I don't know if you are familiar with the
fields of computational linguistics and statistical machine translation. In
statistical machine translation SMT translation patterns are machine learned
from bilingual corpora of translated text. Corpora such as the Europarl
corpus are typically used.

http://www.statmt.org/europarl/

<http://www.statmt.org/europarl/>As you can see in the table of figures
scrolling down a little the number of sentences used in training systems is
in the order of 2 million sentences for each language pair. This probably
seems like a staggering amount of data compared to the size of the data set
of pre-exilic inscriptions. In fact, this data set is too small and
researchers in SMT would like much larger data sets. In general the more
data you throw at the machine learning algorithms the better the translation
models extracted and the better the quality of translations produced by
their systems.

Another vital component of SMT systems is the language model LM. LM's are
n-gram models trained on masses of monolingual data. Obviously, the scope
for the amount of data used in language modelling is much larger because of
the monolingual nature of the data. This means we can trawl the internet for
natural language text of common languages like English. You would think the
internet was large enough to get a good representative LM right? Again,
researchers constantly look for ways to increase the size of their training
data sets because the more data you throw at the learning algorithms the
better the systems work.

Compared to these state of the art technologies your data sets are dwarfed
Yitzhak. There really is no two ways about it. The best you can possibly
hope to achieve is to inspect the data and to make generalisations from the
patterns you observe. The smaller the data set the more likely your
generalisations are to be over generalisations. This is a fact of natural
language.

James Christian

On 25 May 2010 01:47, Yitzhak Sapir <yitzhaksapir AT gmail.com> wrote:

> James, I'm happy to read that you thought over this during the last few
> days.
> Actually, I hoped you were reading up on evidence, but I guess not.  You
> suggest an idea that perhaps the author was trying to omit superfluous
> letters.  Well, there's an easy way to check that.  Let's see if there's an
> inscription that has superfluous letters but defective orthography not in
> the
> Bible.  Yes there is!  Lachish 3 has -th endings for the 2ms suffix, but
> has
> the spelling lhgd for Biblical lhgyd (found in the Bible always with a -y-,
> about 25 times or so), hnb) for BIblical hnby) (found in the Bible over a
> hundred times, also always with a -y-).  It also has )th for Biblical )wtw
> (it,
> 1ms suffixed direct object marker).  Unfortunately, the direct object
> marker
> occurs too many times in the Bible that the concordance I have doesn't list
> them, so I can't tell you how many times it is found with a final -w.  But
> it
> is always spelled with a final -w.
>
> Anyway, James, I think we pretty much have reached a dead end in this
> discussion.
>
> You seem unwilling to look at the evidence, that is, the body of
> pre-exilic epigraphic
> inscriptions.  I am beginning to suspect that perhaps you may be
> trying to filter out
> certain "potentially harmful" information.  This is the only way I can
> figure out to
> explain why we are having a discussion on whether you should spend 10
> minutes
> reading an article that I sent to answer a specific question that you
> asked.  It really
> boggles my mind.
>
> So yes, there is a quick response.  I already told you, "the evidence
> is not just royal
> or official inscriptions."  But then you asked, "Please elaborate if
> this is truthfully and
> demonstrably the case."  So I sent a link to an article that deals
> with the subject.
> Now, you want to go back to the quick response.  I don't understand
> why we're going
> in circles.
>
> James, I recognize that you have a life outside of this list.  I do
> too!  While I enjoy
> discussing Biblical Hebrew, discussing whether or not you are going to read
> an
> article that I found that answers a specific question by you is
> something I'd rather
> not do.  If you don't want to read it, fine!  If you don't want to sit
> down and read the
> epigraphic evidence, that's fine too.
>
> But James, what we need now is not a "linguistic apparatus" or whatever
> other
> buzzword you come up with.  What we need is for you to stop beating around
> the
> bush and contend with the evidence.  Unless you actually EXAMINE the
> epigraphic
> evidence, there's no way that you can reasonably make any kind of statement
> about it regarding the variety, number, breadth, or contents.  You aren't
> even
> in a position to suggest "factors" that may affect the conclusion, nor
> suggestions
> about possible inscriptional standards that were not valid for some reason
> for
> the Bible, or even the Ten Commandments.  It's very simple, a person -- any
> person -- needs to examine the evidence to make reasonable statements
> about it.
>
> And James, if you tell me the local libraries you have access to, I'll
> help you find
> a handbook of inscriptions.  After you read through it, we'll be on
> equal footing in
> this discussion.  And, James, you too may be convinced too that the
> spelling of
> the Bible is later than the pre-exilic standards.
>
> Until then, however, we remain with the simple fact that the evidence
> points in
> a very specific way.  All scholars, or at least, the vast majority of
> scholars, of all
> kinds of religious beliefs, accept the conclusion that the spelling of
> the Bible is
> exilic or post-exilic, based on that evidence, if they are familiar with
> it.
>
> Also, until then, and until you decide you're willing to read the answers
> that I
> post to your questions -- even if they are simply articles that I found
> that
> answer your questions beautifully -- we can't continue the discussion.
>
> Yitzhak Sapir
> _______________________________________________
> b-hebrew mailing list
> b-hebrew AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/b-hebrew
>