The problem of proposing 7-bit transliteration schemes corresponding to a scheme with diacritics may be approached by the concept of transliteration circles.
A transliteration circle is a set of latin characters with diacritics, such that the 7-bit treatment of any given member of the group affects the treatment of at least one other member. The tables below clarify this definition. Abbreviations [some shorter than usual]: -a above br breve gr grave rg ring -b below cb candrabindu lig ligature tl tilde ac acute dt dot mc macron und underline [-a is the default]
Example 1: Since n_tl traditionally becomes ~n, the ~ cannot be used for any other case of n+diacritic, and all these "n's" are in the same circle.
Example 2: If dt-b is going to be treated uniformly, all cases of dt-b are in the same circle.
Example: four transliteration circles:
Circle 4 is arranged so that rows contain the same diacritics and columns
the same latin letters. Hence
- diacritics in the same row may have the same transliteration. - diacritics in the same column must have different transliterations (except in special cases).
(Limited space makes the last column a conflation of four columns!). The first column shows serial no. and any traditional 7-bit transliteration.
Notation: U means u, e, or o; V means any vowel.
t/d_dt-b means t_dt-b and d_dt-b; etc.
mc-a is usually transliterated either by doubling or by using upper case, so it forms a separate transliteration circle.
----------
1. _mc
----------
2. ae_lig
----------
3. c_circ
-------------------------------------------------------------------------
4.
~ V_tl n_tl
. n_dt-b r_dt-b l_dt-b m_dt-b s_dt-b h_dt-b t/d_dt-b
n_mc-b r_mc-b l_mc-b h_mc-b k_mc-b t_mc-b
U_br n_br r_br m_br
n_dt-a m_dt-a y/g_dt-a
r_rg-b l_rg-b
V_ac s_ac
V_gr
m_cb
h_br-b
V_und kh_und
--------------------------------------------------------------------------
(In circle 4, the remaining vowels which can appear with _tl, _ac, or _gr do not affect the scheme, except that elegance might influence the accented combinations _rg-b_ac, _rg-b_gr.)
To make a scheme, two or more rows may be combined, provided there is space available for each character (except in special cases).
If it is decided that some characters with the same diacritic are to be treated differently from the others, then they are to be moved to a separate row.
Proposed scheme
The specification is: case-insensitive, highly readable, modifiers to
come before the letter modified (except for 'h' indicating aspiration),
modifiers to be restricted to the six characters ; , . ~ ^ _
The transliteration circle 4 has 11 rows but there are only 6 allowed modifiers. We move the U_ac/s_ac row to the _dt-a row, U_gr and kh_und to the _mc-b row, m_cb to the _tl row, and h_br-b to the _br row.
Here kh_und is taken to be a special case. Its initial 'k' will be preceded by the same modifier as k_mc-b, but is distinguished by its final 'h'.
Finally, since y/g_dt-a are likely to look a little heavy, they are moved to the _dt-b row.
The resulting scheme is defined as follows (t/d_dt-b follow the regular pattern for _dt-b):
Note: V means any vowel
--------------------------------------------------------------------
1. doubling _mc [but ae-lig_mc is aee, r_rg-b_mc is ,rr]
--------------------------------------------------------------------
2. ae_lig = ae
--------------------------------------------------------------------
3. c_circ = ^c
--------------------------------------------------------------------
4.
~ V_tl n_tl m_cb
. n_dt-b r_dt-b l_dt-b m_dt-b s_dt-b h_dt-b y/g_dt-a
_ V_und n_mc-b r_mc-b l_mc-b h_mc-b (k_mc-b t_mc-b
(kh-und
^ U_br n_br r_br m_br h_br-b
; V_ac n_dt-a m_dt-a s_ac
, V_gr r_rg-b l_rg-b
---------------------------------------------------------------------
Note: if ,r+r occurs it is written ,r:r
Copyright (C) Anthony P. Stone 1997. This material may be freely used,
provided the author is acknowledged
Last updated: 10 June 2002