=====Transliterative Alphabet=====

In the interest of consistency, it is our aim to ensure that each Hebrew word is interpreted as a single word with a single spelling. Only in this manner can an accurate analysis of any given word occur.

To this end we have developed a bilaterally transliterative alphabet which has a one-to-one correspondence in both Hebrew and English, and as a side benefit, is relatively readily pronounceable to an English speaker, without straying too far from the probable Hebrew pronunciation.

We use this alphabet for dictionary entries, so that they are searchable using an English keyboard. It may be helpful to think of these transliterated words as mnemonics rather than being purely translative or representational.

(This alphabet is not intended to address any inherent pronunciation issues that exist withing the Hebrew language, whether modern or ancient. Nevertheless, we refer to "sounds" and "pronunciations" here as a sort of shorthand for spelling standards.)

With this bilateral alphabet it is possible to flip back and forth between Hebrew and English with single characters, thus avoiding two common problems with Hebrew transliteration:
  -the problem of **multiple** (non-unique) Hebrew letters that map to a single English sound
  -the problem of **complex** Hebrew sounds that map to character pairs such as "sh", "ch", "tz"

Consider an example containing both problems. The Hebrew letter צ is pronounced "tz". It is a complex sound. On top of this, the "t" sound is used for two Hebrew letters; ת and ט. Therefore, the word pronounced 'itza' does not uniquely map back to Hebrew, but instead can be correctly mapped back three different spellings: יתזא, יטזא, or יצא. Clearly this is problematic for translators as it is impossible to know which Hebrew word is referred to. 

Both of these problems can be avoided by utilizing a one-to-one character mapping.

^Hebrew	^Transliteration	^Pronunciation|
|א‎	|a|	a|
|ב‎	|b|	b|
|ד‎	|d|	d|
|ג‎	|g|	g|
|ה‎	|h|	h|
|י‎	|i|	i|
|כ‎	|k|	k|
|ל‎	|l|	l|
|מ‎	|m|	m|
|נ‎	|n|	n|
|ו‎	|o|	o|
|פ‎	|p|	p|
|ק‎	|q|	k|
|ר‎	|r|	r|
|ס‎	|c|	s|
|ש	|s|	sh|
|[[ט‎]]	|f|	t|
|ת‎	|t|	t|
|צ‎	|x|	tz|
|ע‎	|y|	y|
|ז‎	|z|	z|

Most of the transliterated letters are pronounced intuitively for the English speaker, but a few minor caveats should be noted:
  * both כ‎ and ק‎ map to a "k" sound. The transliteration of כ‎ is 'k' and of ק‎ is 'q'
  * both ש‎‎ and ס‎‎ map to a "s" sound, although sometimes ש‎ tends to shift toward "sh". The transliteration of ש‎‎ is 's' (based on phonetic /ʃ/) and of ס‎‎ is 'c'. (think of a soft c as in cycle or cent)
  * both ת‎‎ and [[ט‎‎]] map to a "t" sound. The transliteration of ת‎‎ is 't' and of ט‎‎ is 'f'. The letter f was chosen because it was unused, and vaguely resembles an upside down 't'. It should not me mistaken for פ‎ which sometimes is pronounced "f" in other Hebrew transliteration schemes.
  * the letter צ‎ is pronounced "tz". The chosen transliteration of צ‎ is 'x'

We acknowledge that this choice comes with a moderate level of degeneration to the level of pronunciation usability, and consider the increase in written clarity a higher priority.

This alphabet also clarifies the use of [[vowels]] in Hebrew, by highlighting that all vowel sounds are present natively without any requirement for [[diacritics]]. The exception is the letter "e" which is simply inserted as a default in the lack of any other explicit vowels ("a", "i", or "o")

====Usage Notes====
  *All consonants are followed by a vowel unless at the end of a sentence. The default vowel is "e" unless otherwise specified
  *the process for transliterating from Hebrew is as follows:
    - ignore [[diacritics]] and [[affixes]]
    - map all letters using the chart above
    - add "e" after each consonant that is not already followed by another vowel, unless it is an "h" at the end of the word which essentially functions as a vowel.
      * an ending ה‎ results in an "eh" sound, unless it is immediately preceded by a vowel, in which case it modifies the vowel's normal sound as in "ah", "ih", or "oh"  
    - map pronunciation from the transliteration as required for ק‎(q->k), ס(c->s), ש(s->sh), ט‎(f->t), צ(x->tz)
  *any transliterated word can be mapped back to Hebrew by a two step process:
    - ignore all "e"s
    - map all other letters using the chart above

====Shortcomings and Problems====
There are several known shortcomings inherent in this transliterative alphabet system.

As with any linear algebra mapping, it is desirable for a transformation system to be both one-to-one and onto, that is to say that a mapping across the transform yields one and only one corresponding solution. 
In order for our transliteration system to meet these criteria it is necessary to officially exclude the vast numbers of alternate pronunciations and alternate spellings that exist in the current Hebrew environment (particularly but not limited to modern interpretations). We can easily prove that it does not entirely map onto the set of (Modern Hebrew + Ancient Hebrew), but the hope is that the criteria can be met for the limited subset of a more standardized spelling of the ancient Hebrew lexicon only. We will not be able to test this hypothesis until every word in the lexicon has been run through our transliteration algorithm.

Certain scenarios exist where the possible pronunciations allowed within the system appear to fall short of actual pronunciations used in the verbal language. However, the intention of this mapping is to cover Ancient Hebrew only. We are not attempting to cover Modern Hebrew. Thus, it is important to differentiate between the Ancient and Modern versions; a task which is stymied by the fact that only the Modern pronunciations can truly be known. We are forced to perform some guesswork when it comes to the pronunciation of the Ancient Hebrew, based on the extant written evidence and informed (but not necessarily limited) by the Modern pronunciations, even though we know there to be significant differences.

Ultimately though, the goal of this transliterative alphabet is not to provide consistent [[pronunciation]] but rather, to provide a tool capable of aiding in the discovery of [[semantic consistency]] within the Ancient corpus.

Since the discussion has necessarily turned to linear algebra, it should be stated also that shortcomings are not one-to-one and onto with problems. That is to say, that  the admitted shortcomings reveal only //potential// problems. They may or may not materialize into actual problems, dependent upon how they affect our ability to achieve our goal of [[semantic consistency]]. Each case must be evaluated by its specific impact. So what are the possibly problematic cases?

  * leading and trailing [[a vs e]] are indistinguishable
  * the "J" sound appears to be represented in the verbal but not the written language. This is explored further under the heading: [[#The Judah Problem]]
  * the "V" sound appears to be represented in the verbal language but is inadequately represented in the written language. This is explored further under the heading: [[#The Levi Problem]].
    

Other strange cases:
  * [[#the "Sabbath" problem]](s)
  * a few other weird pronunciation topics

Each of this cases will be assumed to be problematic, then examined to determine whether any real problem exists in terms of our stated goal.

===the Judah problem===
The word "Judah" is exemplary of a number of issues. In Hebrew, it is spelled יהודה. We transliterate this into '[[ihodeh]]'. Exactly how the "ih" pattern becomes a "j" sound is unclear. We do not know if the "j" sound was present in the ancient oral language, or if it developed later. 
Other words (for example 'gihon') use the same "ih" letter pattern and do not imply a "j" sound.

A secondary problem exists in the word "Judah". How the "o" turned into a long "u" sound. It should be noted that even in English an o often takes on this sound, although we are accustomed to double the letter visually to form "oo".

  *the "j" sound appears to be represented in the verbal language but is inadequately represented in the written language. Did this sound exist in the ancient language?
  *the "u" (or "oo") sound appears to be represented in the verbal language but is inadequately represented in the written language. Did this sound exist in the ancient language?
===the Levi problem===
The word לוי transliterates to 'loi'.

At first glance this may appear to be an example of a rather common occurrence of the Hebrew letter ו seeming to perform multiple duties. It happens quite regulary that the ו, though generally sounding like "o", will often take on a trailing consonant sound and become "ov". That may explain the presence of the "v" sound. But does it? If this were the case, we would have "lovi", not "levi". 

Strangely, a "v" sound often occurs in common Hebrew pronunciations, where we should expect a "b". That is to say, the letter ב is usually understood to correspond to "b", but sometimes it is a "v" instead. This is generally explained away with the argument that the diacritic dot in the middle of בּ turns a "b" into a "v". A valiant effort by it's proponents, but we have already addressed how trustworthy [[diacritics]] are.

The usual pronunciation of "levi" emphasises the long e sound, which we would normally spell using "ee". However, this is an additional problem. The Hebrew alphabet does not include such a sound. The closest we get would be a short 'i' which is represented by י. But in this particular word, that letter appears only at the end, and is supposedly pronounced as a long i, something fairly atypical for the letter י.

  *the "v" sound appears to be represented in the verbal language but is inadequately represented in the written language. Did this sound exist in the ancient language?
  *the "ee" appears to be represented in the verbal language but is inadequately represented in the written language. Did this sound exist in the ancient language?
  *the "long i" sound appears to be represented in the verbal language but is inadequately represented in the written language. Did this sound exist in the ancient language?


===the many sounds of "o"===
FIXME
o oo/u v

sample data:
  * 2341 hoileh - havilah

then theres the other side of the coin - non O pronounces as O
  * find an example with y - o
  * a - o
  * null as o - see occasional alternate spellings of [[gebor]], and [[neheset]]

===the Sabbath problems===
FIXME "shabbat" or "shabbos"?

double "b"

where's the "th"

===the Cherubim problem===
FIXME
===the Nephilim problem===
FIXME
===Leading E===
leading [[a vs e]] are indistinguishable
FIXME
===Trailing E===
trailing [[a vs e]] are indistinguishable
FIXME

===Sometimes Y===

Consider the word גלעד. It is pronounced "Gilead", and this pronunciation is very closely aligned with the transliteration of '[[geleyed]]'. In this case it is clear that the letter ע is functioning as a consonant, as in the English word 'yes', rather than a vowel, and that the vowel sounds are supplied by the "invisible E". Although there are probably other examples where ע may have been intended to act more like a vowel, it is impossible to know for sure, and our approach is to treat the letter ע consistently.

As another example, consider another well known word, בעל, which is typically rendered "Baal" and pronounced with a hard A very similar to "bail" or "bale". We see that English has two different words with different spellings that match up to this pronunciation. Other possible spelling are possible as well, such as "bayel" or even "beyel". It is the latter which matches our chosen orthography, and in fact is the exact transliteration our algorithms output for בעל - [[beyel]]. 

===Ghost Vowels===
example [[pix]], [[gedol]]