[apple/icu.git] / icuSources / data / translit / si_si_FONIPA.txt

# © 2016 and later: Unicode, Inc. and others.
# License & terms of use: http://www.unicode.org/copyright.html#License
#
# File: si_si_FONIPA.txt
# Generated from CLDR
#

# Sinhala pronunciation rules
#
# Output
#     k ɡ ŋ ᵑɡ c ɟ ɲ ʈ ɖ ⁿɖ t d n ⁿd p b m ᵐb j r l w ʃ s h f
#     ə əː a aː æ æː i iː u uː e eː o oː
#
# References
# [1] Asanka Wasala, Ruvan Weerasinghe, and Kumudu Gamage:
#     Sinhala Grapheme-to-Phoneme Conversion and Rules for Schwa Epenthesis.
#     Proceedings of the COLING/ACL 2006 Main Conference Poster Sessions,
#     pages 890–897. http://www.aclweb.org/anthology/P06-2114
# Simplify ya + yansaya to plain ya after a consonant.
[\u0D9A-\u0DC6] \u0DCA (\u200D)? { ය\u0DCA‍ය → ය;
# Delete ZWNJ and ZWJ to simplify further processing.
\u200C → ;
\u200D → ;
# Insert a schwa after every consonant that is not followed by a dependent vowel
# or virama.
::Null;
([\u0D9A-\u0DC6]) } [^\u0DCA-\u0DDF \u0DF2\u0DF3] → $1 ə;
# Pronunciation rules proper.
::Null;
# fප is an alternative spelling of ෆ.
# This occurs e.g. in ඩේව\u0DD2ඩ\u0DCA කොපර\u0DCAfප\u0DD3ල\u0DCAඩ\u0DCA (David Copperfield)
# [see http://bradshawofthefuture.blogspot.com/2013/02/f.html].
[Ff]ප → f;
# zස is seemingly the only way to unambiguously indicate a voiced /z/ sound.
# This occurs in e.g. ඇල\u0DCAzසය\u0DD2ම' රෝගය (Alzheimer's disease)
# [see https://si.wikipedia.org/wiki/ඇල\u0DCAzසය\u0DD2ම%27_රෝගය]
# or in zස\u0DD3බ\u0DCA‍රා (zebra) [see https://si.wikipedia.org/wiki/‍zස\u0DD3බ\u0DCA‍රා].
[Zz]ස → z;
ං → ŋ;
o → ŋ;  # common substitution for anusvaraya
ඃ ([\u0D9A-\u0DC6]) → | $1 \u0DCA $1;  # TODO: check which consonants geminate
ඃ → h;
අ → a;
ආ → aː;
ඇ → æ;
ඈ → æː;
ඉ → i;
ඊ → iː;
උ → u;
ඌ → uː;
ඍ → ri;
ඎ → ruː;
ඏ → ilu;
ඐ → iluː;
එ → e;
ඒ → eː;
ඓ → aj;
ඔ → o;
ඕ → oː;
ඖ → aw;  # TODO: check if this is correct
ක → k;
ඛ → k;
ග → ɡ;
ඝ → ɡ;
ඞ → ŋ;
ඟ → ᵑɡ;
ච → c;
ඡ → c;
ජ → ɟ;
ඣ → ɟ;
ඤ → ɲ;
ඥ → kɲ;  # TODO: double-check
ඦ → ɟ;
ට → ʈ;
ඨ → ʈ;
ඩ → ɖ;
ඪ → ɖ;
ණ → n;
ඬ → ⁿɖ;
ත → t;
ථ → t;
ද → d;
ධ → d;
න → n;
ඳ → ⁿd;
ප → p;
ඵ → p;
බ → b;
භ → b;
ම → m;
ඹ → ᵐb;
ය → j;
ර → r;
ල → l;
ව → w;
ශ → ʃ;
ෂ → ʃ;
ස → s;
හ → h;
ළ → l;
ෆ → f;
\u0DCA → ;  # delete virama
ා → aː;
ැ → æ;
ෑ → æː;
\u0DD2 → i;
\u0DD3 → iː;
\u0DD4 → u;
\u0DD6 → uː;
ෘ → ru;
ෙ → e;
ේ → eː;
ෛ → aj;
ො → o;
ෝ → oː;
ෞ → aw;  # TODO: check if this is correct
ෟ → lu;
ෲ → ruː;
ෳ → luː;
# Heuristics for turning /ə/ into /a/. Based on [1].
$c=[k ɡ ŋ {ᵑɡ} c ɟ ɲ ʈ ɖ {ⁿɖ} t d n {ⁿd} p b m {ᵐb} j r l w ʃ s z h f];
$s=[:^L:];
# Rule #1
::Null;
$s sv    { ə      → ə;  # exception (a)
$s k     { ə } r  → ə;  # exception (b)
$s $c    { ə } $s → ə;  # exception (c)
$s $c $c { ə      → a;
$s $c    { ə      → a;
# Rule #2
::Null;
$c r { ə } $c → a;  # clause (a) and (b)
$c r { a } h  → a;  # clause (d), exception
$c r { a } $c → ə;  # clause (c)
# Rule #3
# The paper is unclear about what this rule means. The interpretation here
# assumes that "preceded" in the paper is a typo and should be read "followed".
::Null;
[a e æ o ə] h { ə → a;
# Rules #4 through #7
::Null;
ə } $c $c     → a;  # Rule #4
ə } [rbɖʈ] $s → ə;  # Rule #5 exception
ə } $c     $s → a;  # Rule #5
ə } ji     $s → a;  # Rule #6
k { ə } [rl] u    → a;  # Rule #7
# Rule #8
# Note that the paper doesn't say explicitly that this rule should be
# anchored at the beginning of a word, but the remarks before the rules
# seem to imply this.
::Null;
$s k { a } l[aeo]ːj   → ə;  # Typo in paper: /j/ was /y/.
$s k { a } le[mh][ui] → ə;
$s k { alə } h[ui]    → əle;
$s k { a } lə         → ə;
# Diphthongs
::Null;
www+ → ww;  # යෞව\u0DCAවන
[i {iː} e {eː} æ {æː} o {oː} a {aː}] { wu → w;
əji → aj;
iji → iː;  # perhaps: ij
[u {uː} e {eː} æ {æː} o {oː} a {aː}] { ji → j;
Commit	Line	Data
f3c0d7a5 A	1	# © 2016 and later: Unicode, Inc. and others.
	2	# License & terms of use: http://www.unicode.org/copyright.html#License
	3	#
2ca993e8	4	# File: si_si_FONIPA.txt
f3c0d7a5	5	# Generated from CLDR
2ca993e8 A	6	#
	7
	8	# Sinhala pronunciation rules
	9	#
	10	# Output
	11	# k ɡ ŋ ᵑɡ c ɟ ɲ ʈ ɖ ⁿɖ t d n ⁿd p b m ᵐb j r l w ʃ s h f
	12	# ə əː a aː æ æː i iː u uː e eː o oː
	13	#
	14	# References
	15	# [1] Asanka Wasala, Ruvan Weerasinghe, and Kumudu Gamage:
	16	# Sinhala Grapheme-to-Phoneme Conversion and Rules for Schwa Epenthesis.
	17	# Proceedings of the COLING/ACL 2006 Main Conference Poster Sessions,
	18	# pages 890–897. http://www.aclweb.org/anthology/P06-2114
	19	# Simplify ya + yansaya to plain ya after a consonant.
	20	[\u0D9A-\u0DC6] \u0DCA (\u200D)? { ය\u0DCA‍ය → ය;
	21	# Delete ZWNJ and ZWJ to simplify further processing.
	22	\u200C → ;
	23	\u200D → ;
	24	# Insert a schwa after every consonant that is not followed by a dependent vowel
	25	# or virama.
	26	::Null;
	27	([\u0D9A-\u0DC6]) } [^\u0DCA-\u0DDF \u0DF2\u0DF3] → $1 ə;
	28	# Pronunciation rules proper.
	29	::Null;
	30	# fප is an alternative spelling of ෆ.
	31	# This occurs e.g. in ඩේව\u0DD2ඩ\u0DCA කොපර\u0DCAfප\u0DD3ල\u0DCAඩ\u0DCA (David Copperfield)
	32	# [see http://bradshawofthefuture.blogspot.com/2013/02/f.html].
	33	[Ff]ප → f;
	34	# zස is seemingly the only way to unambiguously indicate a voiced /z/ sound.
	35	# This occurs in e.g. ඇල\u0DCAzසය\u0DD2ම' රෝගය (Alzheimer's disease)
	36	# [see https://si.wikipedia.org/wiki/ඇල\u0DCAzසය\u0DD2ම%27_රෝගය]
	37	# or in zස\u0DD3බ\u0DCA‍රා (zebra) [see https://si.wikipedia.org/wiki/‍zස\u0DD3බ\u0DCA‍රා].
	38	[Zz]ස → z;
	39	ං → ŋ;
	40	o → ŋ; # common substitution for anusvaraya
	41	ඃ ([\u0D9A-\u0DC6]) → \| $1 \u0DCA $1; # TODO: check which consonants geminate
	42	ඃ → h;
	43	අ → a;
	44	ආ → aː;
	45	ඇ → æ;
	46	ඈ → æː;
	47	ඉ → i;
	48	ඊ → iː;
	49	උ → u;
	50	ඌ → uː;
	51	ඍ → ri;
	52	ඎ → ruː;
	53	ඏ → ilu;
	54	ඐ → iluː;
	55	එ → e;
	56	ඒ → eː;
	57	ඓ → aj;
	58	ඔ → o;
	59	ඕ → oː;
	60	ඖ → aw; # TODO: check if this is correct
	61	ක → k;
	62	ඛ → k;
	63	ග → ɡ;
	64	ඝ → ɡ;
	65	ඞ → ŋ;
	66	ඟ → ᵑɡ;
	67	ච → c;
	68	ඡ → c;
	69	ජ → ɟ;
70	ඣ → ɟ;
71	ඤ → ɲ;
72	ඥ → kɲ; # TODO: double-check
73	ඦ → ɟ;
74	ට → ʈ;
75	ඨ → ʈ;
76	ඩ → ɖ;
77	ඪ → ɖ;
78	ණ → n;
79	ඬ → ⁿɖ;
80	ත → t;
81	ථ → t;
82	ද → d;
83	ධ → d;
84	න → n;
85	ඳ → ⁿd;
86	ප → p;
87	ඵ → p;
88	බ → b;
89	භ → b;
90	ම → m;
91	ඹ → ᵐb;
92	ය → j;
93	ර → r;
94	ල → l;
95	ව → w;
96	ශ → ʃ;
97	ෂ → ʃ;
98	ස → s;
99	හ → h;
100	ළ → l;
101	ෆ → f;
102	\u0DCA → ; # delete virama
103	ා → aː;
104	ැ → æ;
105	ෑ → æː;
106	\u0DD2 → i;
107	\u0DD3 → iː;
108	\u0DD4 → u;
109	\u0DD6 → uː;
110	ෘ → ru;
111	ෙ → e;
112	ේ → eː;
113	ෛ → aj;
114	ො → o;
115	ෝ → oː;
116	ෞ → aw; # TODO: check if this is correct
117	ෟ → lu;
118	ෲ → ruː;
119	ෳ → luː;
120	# Heuristics for turning /ə/ into /a/. Based on [1].
121	$c=[k ɡ ŋ {ᵑɡ} c ɟ ɲ ʈ ɖ {ⁿɖ} t d n {ⁿd} p b m {ᵐb} j r l w ʃ s z h f];
122	$s=[:^L:];
123	# Rule #1
124	::Null;
125	$s sv { ə → ə; # exception (a)
126	$s k { ə } r → ə; # exception (b)
127	$s $c { ə } $s → ə; # exception (c)
128	$s $c $c { ə → a;
129	$s $c { ə → a;
130	# Rule #2
131	::Null;
132	$c r { ə } $c → a; # clause (a) and (b)
133	$c r { a } h → a; # clause (d), exception
134	$c r { a } $c → ə; # clause (c)
135	# Rule #3
136	# The paper is unclear about what this rule means. The interpretation here
137	# assumes that "preceded" in the paper is a typo and should be read "followed".
138	::Null;
139	[a e æ o ə] h { ə → a;
140	# Rules #4 through #7
141	::Null;
142	ə } $c $c → a; # Rule #4
143	ə } [rbɖʈ] $s → ə; # Rule #5 exception
144	ə } $c $s → a; # Rule #5
145	ə } ji $s → a; # Rule #6
146	k { ə } [rl] u → a; # Rule #7
147	# Rule #8
148	# Note that the paper doesn't say explicitly that this rule should be
149	# anchored at the beginning of a word, but the remarks before the rules
150	# seem to imply this.
151	::Null;
152	$s k { a } l[aeo]ːj → ə; # Typo in paper: /j/ was /y/.
153	$s k { a } le[mh][ui] → ə;
154	$s k { alə } h[ui] → əle;
155	$s k { a } lə → ə;
156	# Diphthongs
157	::Null;
158	www+ → ww; # යෞව\u0DCAවන
159	[i {iː} e {eː} æ {æː} o {oː} a {aː}] { wu → w;
160	əji → aj;
161	iji → iː; # perhaps: ij
162	[u {uː} e {eː} æ {æː} o {oː} a {aː}] { ji → j;
163