rules.txt

# morph-ga.txt: Morphology rules for Irish.
# Copyright (C) 2004-2007 Kevin P. Scannell <kscanne@gmail.com>
#
# This is free software; see the file COPYING for copying conditions.  There
# is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
# PURPOSE.
#
#  with the exception of some of the non-standard verb endings,
#  each rule decreases number of caps or length => recursion will stop.
#
#  each substitution is applied globally (though this is only relevant
#  for the few rules that aren't anchored at ^ or $
#
#  There is no need to encode the regular inflectional morphology of 
#  Irish here because it is already handled by the C++ code that generates
#  the basic lexicon.   So if "imirceacha" is not in the lexicon (it is)
#  "imirceach" won't be either.   On the other hand it can be quite useful
#  to include *derivational* morphology; so if "gaimbíneachas" is not
#  in the lexicon (it is) there is some hope that by stripping the 
#  "eachas" it will find the root "gaimbín".
^([ACEFH-LO-SU-ZÁÉÍÓÚ])([A-ZÁÉÍÓÚ'-]*[A-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	$1\L$2	-1
^B([A-GI-OQ-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	B\L$1	-1
^B'([AEIOUÁÉÍÓÚF][A-ZÁÉÍÓÚ'-]*)$	B'\L$1	-1
^BH([A-EG-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	Bh\L$1	-1
^D([A-SU-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	D\L$1	-1
^D'([AEIOUÁÉÍÓÚF][A-ZÁÉÍÓÚ'-]*)$	D'\L$1	-1
^G([ABD-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	G\L$1	-1
^M([AC-ZÁÉÍÓÚ][A-ZÁÉÍÓÚ'-]*)$	M\L$1	-1
^N([ABCEFH-ZÁÉÍÓÚ-][A-ZÁÉÍÓÚ'-]*)$	N\L$1	-1
^T([A-RT-ZÁÉÍÓÚ-][A-ZÁÉÍÓÚ'-]*)$	T\L$1	-1
^([ACEFH-LO-SU-ZÁÉÍÓÚ])	\l$1	-1
^B([^pPh])	b$1	-1
^Bh([^fF])	bh$1	-1
^D([^tT])	d$1	-1
^G([^cC])	g$1	-1
^M([^bB])	m$1	-1
^N([^dDgG])	n$1	-1
^T([^sS])	t$1	-1
(...)-([A-ZÁÉÍÓÚ])	$1-\l$2	-1
^b'([AEIOUÁÉÍÓÚF])	b'\l$1	-1
^bP([A-ZÁÉÍÓÚ'-]+)$	bP\L$1	-1
^bP	bp	-1
^bhF([A-ZÁÉÍÓÚ'-]+)$	bhF\L$1	-1
^bhF	bhf	-1
^d'([AEIOUÁÉÍÓÚF])	d'\l$1	-1
^dT([A-ZÁÉÍÓÚ'-]+)$	dT\L$1	-1
^dT	dt	-1
^gC([A-ZÁÉÍÓÚ'-]+)$	gC\L$1	-1
^gC	gc	-1
^h([AEIOUÁÉÍÓÚ])([A-ZÁÉÍÓÚ'-]+)$	h$1\L$2	-1
^h([AEIOUÁÉÍÓÚ])	h\l$1	-1
^m'([AEIOUÁÉÍÓÚF])	m'\l$1	-1
^mB([A-ZÁÉÍÓÚ'-]+)$	mB\L$1	-1
^mB	mb	-1
^n([AEIOUÁÉÍÓÚ])([A-ZÁÉÍÓÚ'-]+)$	n$1\L$2	-1
^n([AEIOUÁÉÍÓÚ])	n-\l$1	-1
^nD([A-ZÁÉÍÓÚ'-]+)$	nD\L$1	-1
^nD	nd	-1
^nG([A-ZÁÉÍÓÚ'-]+)$	nG\L$1	-1
^nG	ng	-1
^tS([A-ZÁÉÍÓÚ'-]+)$	tS\L$1	-1
^tS	ts	-1
^t([AEIOUÁÉÍÓÚ])([A-ZÁÉÍÓÚ'-]+)$	t$1\L$2	-1
^t([AEIOUÁÉÍÓÚ])	t-\l$1	-1
# these aren't in aspell db by default; better to do early (mb'amhlaidh)
^mb'	b'	-1
^d'f([aeiouáéíóú])	d'fh$1	2
#  do these early - avoids mistakes with "aighthe" (==aghaidheanna)
#  and note that first rule must precede second to get "-aithe" pp's correct
# IMPORTANT - used for -ughadh endings e.g., -> -aghadh which goes to ú below 
u(i?[dg]h)	a$1	1
i[dg]h(th?ear)$	ítear	1
i[dg]h(th?í)	ítí	1
i[dg]hth?(.)	ith$1	1
idhtheach	íoch	1
i[dg]hth?(.)	it$1	1
#######################
#  rules-local.txt here...
(...)-sin$	$1_sin	1
(...)-se?o$	$1_seo	1
^(ao|é[ai])n-(...)	aon_$2	1
^m'([bcdglmnpst])	mo_$1	1
# tsórt -> sórt, tsaghas -> saghas
# also means we need to have forms like "tsaoghail", where we want to
# preserve the initial t, in the IG db as alt
^ts	s	1
# note next two aren't anchored at start => multiple digits are ok 
([0-9])adh$	$1ú	1
([0-9])mhadh$	$1ú	1
e[oó][cg]had$	eoidh_mé	1
([^e])[oó][cg]had$	$1óidh_mé	1
e[oó][cg]hadsa$	eoidh_mise	1
([^e])[oó][cg]hadsa$	$1óidh_mise	1
e[oó]chair$	eoidh_tú	1
e[oó]chairse$	eoidh_tusa	1
(..)eabhair$	$1_sibh	1
(..[^e])abhair$	$1_sibh	1
([^e])[oó]chair$	$1óidh_tú	1
(..)fad-sa$	$1faidh_mise	1
(..)fadsa$	$1faidh_mise	1
(..)fad$	$1faidh_mé	1
(..)fead-sa$	$1fidh_mise	1
(..)feadsa$	$1fidh_mise	1
(..)fead$	$1fidh_mé	1
(..)fai[rs]$	$1faidh_tú	1
(..)fi[rs]$	$1fidh_tú	1
(..)fai[rs]se$	$1faidh_tusa	1
(..)fi[rs]se$	$1fidh_tusa	1
(...)óis$	$1óidh_tú	1
(...)eois$	$1eoidh_tú	1
(...)ód$	$1óidh_mé	1
(...)eod$	$1eoidh_mé	1
(...)óid$	$1óidh_siad	1
(...)eoid$	$1eoidh_siad	1
(..)ains$	$1ann_tú	1
(..[^a])ins$	$1eann_tú	1
(..)íns$	$1íonn_tú	1
(..)[au]íd$	$1aíonn_siad	1
(.[^f])eas$	$1_mé	1
(..[^aeiouáéíóúf])as$	$1_mé	1
(.[^f])eas-sa$	$1_mise	1
(..[^aeiouáéíóúf])as-sa$	$1_mise	1
álas$	áil_mé	1
(...)íos$	$1igh_mé	1
ígheas$	igh_mé	1
(..[^f])ai[rs]$	$1_tú	1
(..[^aeiouáéíóúf])i[rs]$	$1_tú	1
(...)ís$	$1igh_tú	1
#############################################################################
################## CAIGHDEÁN OIFIGIÚIL SPELLING REFORMS #####################
#############################################################################
#   modernize prefixes (no need to trap mutated versions, stripped below
#   full list from OD77 is in gaeilge/diolaim/x/OD77alt-prefixes
^(h?a)ith	$1th	1
^d'aith	d'ath	1
^(h?a)ird([^e])	$1rd$1	1
# handled by more general "nn" rule below
#^h?anns			ans		1	<.+>	<&>
^([dm])h'	$1'	1
^h-	h	1
^n-([AEIOUÁÉÍÓÚ])	n$1	1
^t-([AEIOUÁÉÍÓÚ])	t$1	1
^(h?a)oin	$1on	1
^(h?a)na-	$1n-	1
#^[Áá]rd(-?..)	ard$1		1	<.+>	<&>
^(h?a)th-?th	$1t	1
^bh'	b'	1
^mbh	mb	2
^(m?bh?)ain([^_].)	$1an$2	1
^(m?bh?)áin	$1án	1
^(m?bh?)airr	$1arr	1
^(m?bh?)aoith	$1aoth	1
^(m?bh?)eig	$1eag	1
^(m?bh?)éil	$1éal	1
^(m?bh?)ioth	$1ith	1
^(m?bh?)oig	$1og	1
^(m?bh?)oinn	$1onn	1
^(m?bh?)oirr	$1orr	1
# handled by more general ll rule below
#^bolls			bols		1	<.+>	<&>
^(m?bh?)uadh	bua	1
^(m?bh?)uain	buan	1
^(g?ch?)aim	$1am	1
# handled by more general tht rule below
#^caitht			cait		1	<.+>	<&>
^(g?ch?)aoimh	$1aomh	1
# N.B. makes string longer
#^ceathar			ceathair		1	<.+>	<&>
^(g?ch?)éid	$1éad	1
# N.B. makes string longer
#^cinn			ceann		1	<.+>	<&>
^(g?ch?)einn	$1eann	1
^(g?ch?)ionn	$1eann	1
^(g?ch?)láir	$1lár	1
^(g?ch?)laoidh	$1loí	1
^(g?ch?)laoin	$1laon	1
^(g?ch?)oilg	$1olg	1
^(g?ch?)[oó]imh-mh	$1óimh	1
^(g?ch?)oimh	$1omh	1
^(g?ch?)[óo]imhmh	$1óimh	1
^(g?ch?)óimh([^eéií])	$1omh$2	1
# general ómh[^a] -> omh$1  rule below
^(g?ch?)ómha	$1omha	1
^(g?ch?)ómhmh	$1ómh	1
^(g?ch?)omh-(mh[aáoóuú])	$1ó$2	1
^(g?ch?)omh-(mh[eéií])	$1ói$2	1
^(g?ch?)oimh-n	$1óin	1
^(g?ch?)[oó]mh-?(n[aáoóuú])	$1ó$2	1
^(g?ch?)[oó]mh-?(n[eéií])	$1ói$2	1
^(g?ch?)ó-(..)	$1omh$2	1
^(g?ch?)ó([bcdfgpt]h)	$1omh$2	1
^(g?ch?)[ou]ir([pr])	$1or$2	1
^(g?ch?)rainn	$1rann	1
^(g?ch?)rioth	$1rith	1
^(g?ch?)roim	$1rom	1
^(g?ch?)ruai?dh	$1rua	1
^(g?ch?)úil	$1úl	1
^(n?dh?)á-	$1é	1
^(n?dh?)aill	$1all	1
^(n?dh?)aoir([^s])	$1aor$2	1
#  first spelling of dea- is from na Grianna
^(n?dh?)éagh-	$1ea-	1
^(n?dh?)e[áai]gh-	$1ea-	1
^(n?dh?)e[áai]gh([^-])	$1ea-$2	1
^(n?dh?)eir([bg])	$1ear$2	1
^(n?dh?)ío-	$1í-	1
^(n?dh?)iubhr	$1iúr	1
^(n?dh?)lighi?	$1lí	1
^(n?dh?)oi-	$1o-	1
^(n?dh?)oinn	$1onn	1
^(n?dh?)roich	$1roch	1
^(n?dh?)ruim	$1roim	1
^(n?dh?)uibh	$1ubh	1
^(n?dh?)ubh-	$1ú	1
# handled by more general prefix+gc rule below
#^éagc			éag		1	<.+>	<&>
#^éigc			éig		1	<.+>	<&>
# handled by more general prefix+dt rule below
#^éadt			éad		1	<.+>	<&>
#^éidt			éid		1	<.+>	<&>
# handled by more general ll rule below
#^falls			fals		1	<.+>	<&>
^eadar	idir	1
^headar	hidir	1
^eidir	idir	1
^heidir	hidir	1
^((bh)?fh?)ágbh	$1ág	1
^((bh)?fh?)iadh	$1ia	1
^((bh)?fh?)ír-	$1íor	1
^((bh)?fh?)lith	$1rith	1
^((bh)?fh?)rioth([^á])	$1rith$3	1
^((bh)?fh?)ó-	$1o-	1
^((bh)?fh?)oir-	$1or-	1
^((bh)?fh?)uair-	$1uar-	1
# N.B. makes string longer
#^gamh			geamh		1	<.+>	<&>
# N.B. makes string longer
^(n?gh?)irr	$1earr	1
^(n?gh?)airbh	$1arbh	1
^(n?gh?)éir	$1éar	1
^(n?gh?)náith	$1náth	1
^(n?gh?)obh	$1abh	1
^héala([^aeiouáéíóú]+[eéií])	héili$1	1
^héala([^aeiouáéíóú]+[aáoóuú])	héilea$1	1
^(h?i)air	$1ar	1
^(h?i)arann	$1arn	1
^(h?i)odar	$1dir	1
^(h?i)ol([^ar])	$1l$2	1
^(h?i)omshlá	$1omlá	1
^(h?i)om([^aálpr])	$1m$2	1
^(h?i)on	$1n	1
# handled by more general nn rule below
#^h?ionn([rs])			ion$2		1	<.+>	<&>
^([ls])áimh	$1ámh	1
^láin	lán	1
# handled by more general sb rule below
#^léasb			léasp		1	<.+>	<&>
^leig([^h])	lig$1	1
# risky?
^leith-	leath-	1
^loim	lom	1
^lóir	leor	1
^luaith([^r])	luath$1	1
# handled by more general sg rule below
#^luasg			luasc		1	<.+>	<&>
# N.B. makes string longer
^lubh	luibh	1
^lui([mn])	loi$1	1
^lúith	lúth	1
^(mh?)aill	$1all	1
^(mh?)aoil	$1aol	1
^(mh?)aoith([^n])	$1aoth$2	1
^(mh?)airbh	$1arbh	1
# only two math- words anyway?
#^magh		math		1	<.+>	<&>
^(mh?)eadh[oó]n	$1eán	1
^(mh?)ean	$1eán	1
^mh'([aeiouáéíóú].)	m'$1	1
# N.B. makes string longer
^(mh?)in-	$1ion-	1
^(mh?)ío-	$1í-	1
^(mh?)i-	$1í-	2
^(mh?)ío([bcdfgmpst]h)	$1í$2	1
^(mh?í)-([^aeiouáéíóú].)	$1$2	1
# no midh- words anyway?
#^miodh		midh		1	<.+>	<&>
^(mh?)óir	$1ór	1
^naoimh	naomh	1
^neamh-([^m].)	neamh$1	1
^neimh(..)	neamh$1	1
^nea-mb	neamhbh	1
^nea-	neamh-	1
^nea([cdfgpt]h)	neamh$1	1
^nuadh	nua	1
^(h?ó)i([gr])	$1$2	1
# handled by more general nn rule below
#^pannc			panc		1	<.+>	<&>
#^poinnt			point		1	<.+>	<&>
^(b?ph?)ont	$1unt	1
^(b?ph?)rimh	$1ríomh	2
^(b?ph?)rímh	$1ríomh	1
^réidh	ré	1
^réimh	réamh	1
^riaghal	rial	1
^r[íi]o?gh([^n])	rí$1	1
^ro-	ró-	1
^rói-	ró-	1
^roighn	righn	1
^roimh	réamh	1
^ruadh	rua	1
^rúin	rún	1
^(sh?)áir	$1ár	1
^(sh?)aoghl	$1aol	1
^(sh?)aoibh	$1aobh	1
^(sh?)aoir-	$1aor-	1
^sclug	sclog	1
^scoith	scoth	1
^(sh?)eana-	$1ean-	1
^(sh?)ein-	$1ean-	1
^(sh?)eana([dst])h	$1ean$2	1
^(sh?)iod	$1ead	1
^(sh?)íodh	$1í	1
# N.B. makes it longer
^(sh?)ír-	$1íor-	1
^shm	sm	1
^smúid	smúit	1
^(sh?)oi-	$1o-	1
# handled by more general ll rule below
#^soills			soils		1	<.+>	<&>
# handled by more general nn rule below
#^sonnr			sonr		1	<.+>	<&>
^(sh?)r[áa]ic	$1rac	1
# N.B. makes string longer
^(sh?)rang	$1reang	1
^(sh?)ream	$1ram	1
# N.B. makes string longer - risky, plus no "sruim-" words at all?
#^srim			sruim		1	<.+>	<&>
^stáit	stát	1
^(d?th?)aidhbh	$1aibh	1
# handled by more general sb rule below
#^taisb		taisp	1	<.+>	<&>
^t'([AEIOUÁÉÍÓÚaeiouáéíóú].|[Ff]h)	d'$1	1
^(d?th?)easb([^h])	$1aispe$2	1
^(d?th?)eint	$1int	1
^(d?th?)ele([^aeiouáéíóú]+[eéií])	$1eili$2	1
^(d?th?)ele([^aeiouáéíóú]+[aáoóuú])	$1eilea$2	1
^(d?th?)iom	$1im	1
^(d?th?)iugh	$1iú	1
# risky
#^tiúr	tabhar		1	<.+>	<&>
# handled by more general rule below
#^treabhth			treaf		1	<.+>	<&>
^(d?th?)reas	$1ras	1
^(d?th?)réin	$1réan	1
^(d?th?)roim	$1rom	1
^(d?th?)uaith	$1uath	1
^(h?u)aith	$1ath	1
^(h?ú)ir	$1r	1
# End of Prefixes - Now general spelling reforms (but not from CO handbook)
# Mostly based on corpus work...
# deán for déan
^(n?dh?)eán	$1éan	1
eoracht(a?)$	eoireacht$1	1
óracht(a?)$	óireacht$1	1
éair$	éir	1
eor$	eoir	1
# 08-11-05, generalize nouns to everything (fraochta e.g.)
# Sept 2017, generalize from end of word to anywhere
chd([^h])	cht$1	1
chd$	cht	1
ínteacht(a?)$	íneacht$1	1
óanna$	ónna	1
# corpus - Dómhnach, etc., but not rómhilis, etc.
#  Handles  all ^cómh- prefixes too except  ^cómha which we do above explicitly
([^Rr])ómh([^a])	$1omh$2	1
^(.)ógh	$1ogh	1
eamhn	iún	1
([^e])amhn	$1ún	1
#   warning - makes longer 
^b'f([aeiouáéíóú])	b'fh$1	1
# lán-dhícheall -> lándícheall, sean-shaoghal -> seansaoghal ->  seansaol
([dntls])-([dntls])h	$1-$2	1
#   Now various general spelling reforms from CO handbook
ai(cht?ín)	a$1	1
aa	á	1
sg	sc	1
sb	sp	1
sd	st	1
dn	nn	1
aoí	aoi	1
ín([gs])	in$1	1
éaracht	éireacht	1
(..)[au]ig$	$1aigh	1
ngc	nc	1
é(ir[ld])	e$1	1
á(i?r[dlnr])	a$1	1
ó(i?r[dlnr])	o$1	1
ú(i?r[dlnr])	u$1	1
eó	eo	1
áo	ao	1
úa	ua	1
aó	ao	1
eamh[au]il$	iúil	1
eamhla	iúla	1
([^e])amh[au]il$	$1úil	1
([^e])amhla	$1úla	1
adht([aá])	at$1	1
iomh$	eamh	1
aíle$	aíola	1
idhibh$	í	1
([ií])[dg]hthibh$	$1the	1
aighibh$	aigh	1
([^a])ighibh$	$1igh	1
eachaibh$	igh	1
([^e])achaibh$	$1aigh	1
([^a])chaibh$	$1aigh	1
anaibh$	anna	1
ánaibh$	áin	1
aibh$	a	1
aibh$	aí	1
([^_][^_][^a])ibh$	$1í	1
(.)[ée]u	$1éa	1
tch	t	1
[ií][dg]hea	ío	1
[ií][dg]he	í	1
[ií][dg]hio	ío	1
uí([mto])	aí$1	1
uith	aith	1
adha	ú	1
ádh	á	1
[mb]hth	f	1
th[mb]h	f	1
ghai	aí	1
thch	ch	1
tht	t	1
ll([rst])	l$1	1
nn([cdlrst])	n$1	1
aogha(i?l)	ao$1	1
eamh[au]int$	iúint	1
([^e])amh[au]int$	$1úint	1
([^a])mh[au]int$	$1úint	1
ghail$	aíl	1
ghaile$	aíola	1
ighil$	íl	1
ighile$	íola	1
(ain|[eé][ai]|iar)gc	$1g	1
e[au]gc	éag	1
(é[ai])dt	$1d	1
(dh?í?)omb	$1om	1
(ai?n)bhf	$1bh	1
(..)u(i?s)$	$1a$2	1
#   some dialect/pre-standard noun/adjective inflections
([oó]ga)í$	$1	1
nnaí$	nna	1
(..)dha$	$1tha	1
(..)dha$	$1a	1
eói?g$	eog	1
eoig$	eog	1
óig$	óg	1
lainn$	lann	1
aisí$	ais	1
dth([ea])$	dt$1	1
i[uú]bhar	abhar	1
# teilgioch -> teilgeach, réitioch -> réiteach
ioch$	each	1
# Fórsaíbh -> Fórsaí (dpl)
íbh$	í	1
íocha$	í	1
# but note in previous that sometimes this gets interrupted by correct
# plural adjective:  cogaidheacha->cogaíocha, or sometimes even correct
# plural noun:  claidheacha->claíocha  (so want to keep -íocha intermediate)
#  more patterns from RIA corpus
uía	aío	1
ía	ia	1
^bhfh	bhf	1
^ff	bhf	1
^cc	gc	1
^tt	dt	1
mair$	mar	1
iuc([^h])	ioc$1	1
aé	ae	1
eu	éa	1
i[uú]bha?l	iúl	1
^dfh	d'fh	1
^df([aeiouáéíóú])	d'fh$1	1
##################  NON/PRE-STANDARD VERB ENDINGS #########################
# future/conditional 2nd declension with -cha?- see CO p.103
e[oó]cha$	eoidh	1
([^e])ócha$	$1óidh	1
e[oó]chf?a(i?dh)$	eo$1	1
([^e])óchf?a(i?dh)$	$1ó$2	1
e[oó][cg]haid$	eoidh	1
([^e])[oó][cg]haid$	$1óidh	1
e[oó]chaimi([ds])$	eoimi$1	1
([^e])óchaimi([ds])$	$1óimi$2	1
e[oó]chas$	eoidh	1
e[oó]cham$	eoimid	1
([^e])ócham$	$1óimid	1
e[oó]cha?muid$	eoimid	1
([^e])ócha?muid$	$1óimid	1
e[oó]chamaoid$	eoimid	1
([^e])óchamaoid$	$1óimid	1
e[oó]ch[au]i(nn|dís)$	eoi$1	1
([^e])óch[au]i(nn|dís)$	$1ói$2	1
e[oó]ch[au]innse$	eoinnse	1
([^e])óch[au]innse$	$1óinnse	1
e[oó]ch(th|f)?[aá]$	eofá	1
([^e])óch(th|f)?[aá]$	$1ófá	1
# cond. aut. below are often -óchthaidhe in prestandard corpus
# but an earlier rule maps -aidhe to -aí, so rule below handles both
e[oó]ch(f|th)?(aoi|a?í)$	eofaí	1
([^e])óch(f|th)?(aoi|a?í)$	$1ófaí	1
(ó|eo)chf?ar$	$1far	1
(ó|eo)chthar$	$1far	1
f[au]i[dg]hear$	far	1
fi[dg]hear$	fear	1
[ií]ghean$	íonn	1
eamaoid$	imid	1
([^e])amaoid$	$1aimid	1
eamuid$	imid	1
([^e])amuid$	$1aimid	1
eamaois$	imis	1
([^e])amaois$	$1aimis	1
([^e])adaois$	$1aidís	1
faidhear$	far	1
# dá n-amharcthá, dá dtugthá, srl
th(eá)$	t$1	1
úghadh	ú	1
aghadh	ú	1
iaghadh	íodh	1
(..)thear$	$1tear	1
(..)t(te?ar)$	$1$2	1
ighdar	íodar	1
[íi]ghim	ím	1
[íi]ghmid	ímid	1
ighamaid	ímid	1
aighid$	aíonn	1
([^a])ighid$	$1íonn	1
ighidh	í	1
ighinn	ínn	1
idíst$	idís	1
imíst$	imis	1
uigh'e$	aithe	1
fiomuid$	fimid	1
# fas/feas/óchas/ós/eos here are future relatives
# as on CB p.145, i.e. not synthetic
fas$	faidh	1
feas$	fidh	1
óchas$	óidh	1
ós$	óidh	1
eos$	eoidh	1
# Common in achtanna to have -fidh,-fid ending on broad stem, lots of 
# other slender/broad mismatches:
([aáoóuú][^aeiouáéíóú]*)fí$	$1faí	1
([aáoóuú][^aeiouáéíóú]*)fidh?$	$1faidh	1
([eéií][^aeiouáéíóú]*)fadh$	$1feadh	1
([eéií][^aeiouáéíóú]*)far$	$1fear	1
([eéií][^aeiouáéíóú]*)tar$	$1tear	1
# see achtanna for íocfaid/íocfa, bainfid/bainfe
fa$	faidh	1
faid$	faidh	1
fe$	fidh	1
fid$	fidh	1
# bainfeam -> bainfimid (via Ruairí Ó hUigín, RIA)
fam$	faimid	1
feam$	fimid	1
# Future/Cond autonomous; unwanted prefix h is common in achtanna
^h([aeiouáéíóú].+fe?ar)$	$1	1
^h([aeiouáéíóú].+fa?í)$	$1	1
# leigiodh -> ligeadh, deiriodh -> deireadh, sgéithfiodh -> scéithfeadh
iodh$	eadh	1
iodar$	eadar	1
#  Thugtaoi in MU44, e.g.
taoi$	taí	1
#  OS88 p.126  present direct relative
anns$	ann	1
íonns$	íonn	1
uíon$	aíonn	1
#  corpus; now handled above
# th(eá|ear)$		t$1		1	<V.*>	<&>
#  now handled by more general -amuid rules above
#famuid		faimid		1	<V p="y" t="fáist">	<&>
#  -idh, -aidh on 1st declension verbs: Pres. Subj. or 2nd pl. imper (-igí) 
#  according to Dinneen.  Some examples seem to be pres. in corpus,
#  e.g. suidhidh, some apparently past e.g. uneclipsed "tigidh"
#  Here we go with subjunctive which is the most common by far
(.[^a])idh$	$1e	1
# (..)aidh$	$1a		1	<V p="y" t="foshuit">	<&>
#  pre-standard texts commonly have "déanfaimíd", "tugaimís", "rachaimíd"...
#(...)imí([ds])$		$1imi$2	1	<V.*>	<&>
# generalize and comment out previous...
mí([ds])$	mi$1	1
# these are sometimes also subjunctives: go mbeannuighidh Dia...
# and at least one corpus example that looks like present tense:
# "Tionntuighidh an peacadh an t-anam..."?
[ií]ghidh$	ígí	1
igidh$	igí	1
mist$	mis	1
#  various -áil  endings should be broad, CO p.120
áilim$	álaim	1
áileann$	álann	1
áilfidh$	álfaidh	1
áilfeá$	álfá	1
áilfear$	álfar	1
áilfí$	álfaí	1
ái(lf?)eadh$	á$1adh	1
ái(lf?)im(i[sd])$	á$1aim$2	1
áiligí$	álaigí	1
áilea([dm]ar)$	ála$1	1
ái(lf?)idís$	á$1aidís	1
áltar$	áiltear	1
álta$	áilte	1
álaithe$	áilte	1
álanna$	álacha	1
^h-?([aeiouáéíóú].+)adh$	n-$1adh	1
^h-?([aeiouáéíóú].+)inn$	n-$1inn	1
^h-?([aeiouáéíóú].+)ann$	n-$1ann	1
#############################################################################
##  SYNTHETIC FORMS
##  more non-standard verbs; these are all examples where the pronoun
##  is bound on the non-standard side, but separate in the standard
##  e.g. "glanfad" = "glanfaidh mé"
##  gramadóir architecture doesn't allow multiwords on the replacement
##  side, so we effectively just drop the pronoun here
##  In the Caighdeánaitheoir, we'll comment these out and put analogues
##  in the file rules-local.txt with the appropriate pronouns...
##############################################################################
## Future forms: see OS88 p.215 for some of these
## Future 2nd p., "r" forms noted on OS88 p.216;
## and cf. "fas", "feas" above which are relatives (not synthetic)
#fad$			faidh		1	<V p="y" t="fáist">	<&>
#fead$		fidh		1	<V p="y" t="fáist">	<&>
#fai[rs]$		faidh		1	<V p="y" t="fáist">	<&>
#fi[rs]$		fidh		1	<V p="y" t="fáist">	<&>
#óis$		óidh		1	<V p="y" t="fáist">	<&>  # salóis = salóidh tú
#eois$		eoidh		1	<V p="y" t="fáist">	<&> # imreois = imreoidh tú
#ód$		óidh		1	<V p="y" t="fáist">	<&>  # salód = salóidh mé
#eod$		eoidh		1	<V p="y" t="fáist">	<&> # imreod = imreoidh mé
## socróid = socróidh siad, dílseoid, etc. are 3rd plural forms in achtanna
#óid$		óidh		1	<V p="y" t="fáist">	<&>
#eoid$		eoidh		1	<V p="y" t="fáist">	<&>
##  habitual present; Learning Irish p. 215
#ains$			ann		1	<V p="y" t="láith">	<&> # glanains = glanann tú
#([^a])ins$		$1eann		1	<V p="y" t="láith">	<&> # brisins
#íns$			íonn		1	<V p="y" t="láith">	<&> # salaíns, coinníns
## pretty rare 3rd person pl.; cialluíd in Achtanna, more or less it?
#(..[au])íd$		$1íonn		1	<V p="y" t="láith">	<&> # cialluíd
## past tenses; broad -s is first person, slender -s is 2nd (OS88 p.215) 
## and slender -r is possible in 2nd person (OS88 note on p.216)
## don't allow preceding f to avoid wasting time with -fas, -feas future
## though beware a couple of rare-ish verbs ending in f (graf, liteagraf)
## ** these same endings are above as direct relative present verbs **
#(..[^f])eas$		$1		1	<V p="y" t="caite">	<&>  # bhriseas
#(..[^aeiouáéíóúf])as$	$1		1	<V p="y" t="caite">	<&>  # ghlanas
#álas$			áil			1	<V p="y" t="caite">	<&>	# shábhálas
#áileas$			áil			1	<V p="y" t="caite">	<&>	# shábháileas
#(...)íos$		$1igh		1	<V p="y" t="caite">	<&> # shalaíos, choinníos (but not for syncopated v2 verbs)
#ígheas$		igh		1	<V p="y" t="caite">	<&> # "d'eirígheas as"
#(..[^f])ai[rs]$		$1		1	<V p="y" t="caite">	<&>   # ghlanais
#(..[^aeiouáéíóúf])i[rs]$	$1		1	<V p="y" t="caite">	<&> # bhrisis
#(...)ís$		$1igh		1	<V p="y" t="caite">	<&> # shalaís, choinnís
##############################################################################
##   start some more radical changes, e.g. stripping prefixes completely
##############################################################################
## hyphened prefixes, an-, dea-
#^h?an-([bcfgmp]h)	$1		0	<[AN].*>	<&>
#^h?an-([bcfgmp][^h])	$1		2	<[AN].*>	<&>
#^h?an-([^bcfgmp][^h])	$1		0	<[AN].*>	<&>
#^h?an-([^bcfgmp]h)	$1		2	<[AN].*>	<&>   # an-dhuine 
#^dea-([bcdfgmpt]h)	$1		0	<[AN].*>	<&>
#^dea-([bcdfgmpt][^h])	$1		2	<[AN].*>	<&>
#^dea-(sh[aeiouáéíóúlnr])	$1	0	<[AN].*>	<&>
#^dea-(s[aeiouáéíóúlnr])	$1		2	<[AN].*>	<&>
#^dea-(s[^aeiouáéíóúlnrh])	$1	0	<[AN].*>	<&>   # dea-scéal
#^dea-([^bcdfgmpst])	$1		0	<[AN].*>	<&>
## other prefixes; most are not independent words (ath, comh, etc.)
## but several are -- these are worth including for several reasons:
## (1) efficiency; morph. analysis happens well before the check for
##  compounds/run-togethers (2) allows for a more refined check for
##  lenition of the prefixed word or other changes (comh-mh, comh->có, etc.)
#^h?ain([deéiílnrst])	$1		0	<[AN].*>	<&>
#^h?ain([bcfgmp]h)	$1		0	<[AN].*>	<&>
#^h?ath([^bcdfgmpst-])	$1		0	<[ANV].*>	<&>
#^h?ath([bcdfgmp]h)	$1		0	<[ANV].*>	<&>
#^h?ath(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^h?ath(s[^aeiouáéíóúlnrh])	$1	0	<[ANV].*>	<&>
#^h?ath-?th		th		1	<[ANV].*>	<&>   # -> "at-" above
#^comh([^bcdfgmnpst-])	$1		0	<[ANV].*>	<&>
#^comh([bcdfgpt]h)	$1		0	<[ANV].*>	<&>
#^comh(sh[aeiouáéíóúlnr])	$1	0	<[ANV].*>	<&>
#^comh(s[^aeiouáéíóúlnrh])	$1	0	<[ANV].*>	<&>
## cómhargadh, cómhalartach; but pre-standard "cómhoibriughadh" too
#^có(mh[aáoóuú])		$1		0	<[ANV].*>	<&>
#^cói(mh[eéií])		$1		0	<[ANV].*>	<&>
#^có(n[aáoóuú])		$1		0	<[ANV].*>	<&>
#^cói(n[eéií])		$1		0	<[ANV].*>	<&>
#^do-([aeiouáéíóú])	$1		0	<[AN].*>	<&>
## ^do([lnr]) BELOW
#^do([bcdfgmpt]h)	$1		0	<[AN].*>	<&>
#^do(sh[aeiouáéíóúlnr])	$1		0	<[AN].*>	<&>
#^do(s[^aeiouáéíóúlnrh])	$1		0	<[AN].*>	<&>
#^do-([aeiouáéíóú])	$1		0	<[ANV].*>	<&>
#^droch([^bcdfgmpst-])	$1		0	<[AN].*>	<&>
#^droch-(ch)		$1		0	<[AN].*>	<&>
#^droch([bdfgmpt]h)	$1		0	<[AN].*>	<&>
#^droch(sh[aeiouáéíóúlnr])	$1		0	<[AN].*>	<&>
#^droch(s[^aeiouáéíóúlnrh])	$1		0	<[AN].*>	<&>
#^fíor([^bcdfgmprst-])	$1		0	<[ANV].*>	<&>
#^fíor-(r)		$1		0	<[ANV].*>	<&>
#^fíor([bcdfgmpt]h)	$1		0	<[ANV].*>	<&>
#^fíor(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^fíor(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^for([^bcdfgmprst-])	$1		0	<[ANV].*>	<&>
#^for-(r)		$1		0	<[ANV].*>	<&>
#^for([bcdfgmpt]h)	$1		0	<[ANV].*>	<&>
#^for(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^for(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^fo-([aeiouáéíóú])	$1		0	<[ANV].*>	<&>
## ^fo([lnr]) BELOW
#^fo([bcdfgmpt]h)	$1		0	<[ANV].*>	<&>
#^fo(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^fo(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^fo-(..)		$1		1	<[ANV].*>	<&>
#^frith([^bcdfgmpst-])	$1		0	<[ANV].*>	<&>
#^fri(t[^h])		$1		0	<[ANV].*>	<&>
#^frith([bcdfgmp]h)	$1		0	<[ANV].*>	<&>
#^frith(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^frith(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^h?iar([^bcdfgmprst-])	$1		0	<[ANV].*>	<&>
#^h?iar-(r)		$1		0	<[ANV].*>	<&>
#^h?iar([bcdfgmpt]h)	$1		0	<[ANV].*>	<&>
#^h?iar(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^h?iar(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
## ilsiamsa, ilsiollach, etc.  no special case
#^h?il([^bcfgmp-]..)	$1		0	<[AN].*>	<&>
#^h?il([bcfgmp]h.)	$1		0	<[AN].*>	<&>
#^h?im([^bcdfghmpst-]..)	$1		0	<[ANV].*>	<&>
#^h?im([bcdfgmpt]h.)	$1		0	<[ANV].*>	<&>
#^h?im(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^h?im(s[^aeiouáéíóúlnrh].)	$1		0	<[ANV].*>	<&>
## ^in([^bcfgmp-]) BELOW
#^h?in-(n)		$1		0	<[AN].*>	<&>
#^h?in([bcfgmp]h..)	$1		0	<[AN].*>	<&>
#^leas-(s)		$1		0	<[AN].*>	<&>
#^leas([aeiouáéíóúd]..)	$1		0	<[AN].*>	<&>
#^leas([bcfgm]h..)	$1		0	<[AN].*>	<&>
#^mí-([aeiouáéíóú])	$1		0	<[ANV].*>	<&>
#^mí([aeiouáéíóú]..)	$1		2	<[ANV].*>	<&>  # míúsáid
#^mí([lnr]..)		$1		0	<[ANV].*>	<&>
#^mí([bcdfgmpt]h.)	$1		0	<[ANV].*>	<&>
#^mí(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^mí(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^neamh-?([aeiouáéíóú])	$1		0	<[ANV].*>	<&>  # neamh-aistear but neamhaithne
#^neamh([lnr])		$1		0	<[ANV].*>	<&>
#^neamh([bcdfgpt]h)	$1		0	<[ANV].*>	<&>
#^neamh-(mh)		$1		0	<[ANV].*>	<&>
#^neamh(sh[aeiouáéíóúlnr])	$1	0	<[ANV].*>	<&>
#^neamh(s[^aeiouáéíóúlnrh])	$1	0	<[ANV].*>	<&>
#^príomh([^bcdfgmpst-])	$1		0	<N.*>	<&>
#^príomh([bcdfgpt]h)	$1		0	<N.*>	<&>
#^príomh-(mh)		$1		0	<N.*>	<&>
#^príomh(sh[aeiouáéíóúlnr])	$1		0	<N.*>	<&>
#^príomh(s[^aeiouáéíóúlnrh])	$1		0	<N.*>	<&>
#^réamh([^bcdfgmpst-])	$1		0	<[ANV].*>	<&>
#^réamh([bcdfgpt]h)	$1		0	<[ANV].*>	<&>
#^réamh-(mh)		$1		0	<[ANV].*>	<&>
#^réamh(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^réamh(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^rí([lnr]..)		$1		0	<[ANV].*>	<&>
#^rí([bcdfgmpt]h.)	$1		0	<[ANV].*>	<&>
#^rí(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^rí(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^ró-([aeiouáéíóú]..)	$1		0	<[AN].*>	<&>
#^ró([aeouáéíóú]..)	$1		2	<[AN].*>	<&>  # not "i" since several actual words in Dinneen start with rói-
#^ró([lnr]..)		$1		0	<[AN].*>	<&>
#^ró([bcdfgmpt]h.)	$1		0	<[AN].*>	<&>
#^ró(sh[aeiouáéíóúlnr])	$1		0	<[AN].*>	<&>
#^ró(s[^aeiouáéíóúlnrh])	$1		0	<[AN].*>	<&>
#^sain([^bcdfgmpst-])	$1		0	<[ANV].*>	<&>
#^sain([bcdfgmpt]h)	$1		0	<[ANV].*>	<&>
#^sain(sh[aeiouáéíóúlnr])	$1		0	<[ANV].*>	<&>
#^sain(s[^aeiouáéíóúlnrh])	$1		0	<[ANV].*>	<&>
#^so-([aeiouáéíóú]..)	$1		0	<[AN].*>	<&>
#^so([lnr]..)		$1		0	<[AN].*>	<&>
#^so([bcdfgmpt]h.)	$1		0	<[AN].*>	<&>
#^so(sh[aeiouáéíóúlnr].)	$1		0	<[AN].*>	<&>
#^so(s[^aeiouáéíóúlnrh]..)	$1		0	<[AN].*>	<&>
#^tras([^bcfghmnps-])	$1		0	<[ANV].*>	<&>   # leave "^trasn" alone
#^tras([bcfgmp]h)	$1		0	<[ANV].*>	<&>
#^tras-(s)		$1		0	<[ANV].*>	<&>
## next batch are non-words so really need to be here, but VERY infrequent
#^(cil|gig|h?is|meig|micr|pic|teil)ea-?([^aeiouáéíóú-]+[aouáóú])	$2	0	 <.+>	<&>
#^(cil|gig|h?is|meig|micr|pic|teil)i-?([^aeiouáéíóú-]+[eiéí])	$2	0	 <.+>	<&>
#^(h?ant|fót|nan|par|pol|h?ultr)a-?([^aeiouáéíóú-]+[aouáóú])		$2	0	 <.+>	<&>
#^(h?ant|fót|nan|par|pol|h?ultr)ai-?([^aeiouáéíóú-]+[eiéí])		$2	0	<.+>	<&>
#^(h?eachtar|freas|h?íos|neas|réalt|tob|h?uas|h?uath)-?([^-])		$2	0	 <.+>	<&>
## END PREFIX STRIPPING
#############################################################################
#  Spelling standardizations, but ones where I'm less sure that they 
#  always apply, i.e. non-standard bits that are also commonly standard,
#  e.g. -thaí$   which could be legit "torthaí" but also non-std "glórthaí"
#  Because of this, safer to apply them after the prefix stripping, so
#  e.g. "iarchoirpigh" comes out as MOIRF{choirpigh} instead of 
#  CAIGHDEAN{choirpeach}, and "fíorghanntanas" comes out as 
#  MOIRF{ghanntanas} instead of CAIGHDEAN{ghanntan}
chaí$	cha	1
thaí$	tha	1
(.[aá])igh$	$1ch	1
(.[^aá])i[dg]h$	$1each	1
(..)aidh$	$1a	1
#  rule for [^eéií] is above with other non-std prefixes, last resort here
#  for words like cóimhéigean -> comhéigean
^cóimh([eéií])	comh$1	1
#  Next, some verb endings that I wanted to move *after* their
#  past-tense friends since these are less likely analyses... 
#  CB p.145, Dinneen, PDF Christian Brothers 14.19; direct rel. present
#  exact same endings in the past tense below; in that case, synthetic 1st sing
(..[^f])eas$	$1eann	1
(..[^aeiouáéíóúf])as$	$1ann	1
áileas$	álann	1
(...)íos$	$1íonn	1
ígheas$	íonn	1
#############################################################################
## start derivational morphology
#([ao])chai?s$		$1ch		0	<A pl="n" gnt="n".*>	<&>
#acht$			ach		0	<A pl="n" gnt="n".*>	<N pl="n" gnt="n" gnd="f">
#([ao])chtaí$		$1cht		1	<N pl="n" gnt="n".*>	<&>
#íocht$			íoch		0	<A pl="n" gnt="n".*>	<N pl="n" gnt="n" gnd="f">
#úlachta?$		úil		0	<A pl="n" gnt="n".*>	<&> # CO p.120
#úchái?n$		ú		0	<N.*>	<&>
## but actually non-standard if it's gs or pp; general  uith -> aith below
#uithe$			ú		0	<N.*>	<&>
#([eiéí][^aeiouáéíóú]+)íní?$	$1		0	<.+>	<&>	# diminutive
#([aouáóú][^aeiouáéíóú]+)íní?$	$1		2	<.+>	<&>	# diminutive
#([óo])igíní?$		$1g		0	<N.*>	<&>
##  end derivational morphology
#############################################################################
##  start emphatic suffixes
#(.[aouáóú])(san?|na)$	$1		0	<[NV].*>	<&>
#(.[eiéí])(sean|[sn]e)$	$1		0	<[NV].*>	<&>
#([aouáóú][^aeiouáéíóú-]*[^aeiouáéíóús-])san?$	$1	0	<[NV].*>	<&>
#([aouáóú][^aeiouáéíóú-]*[^aeiouáéíóún-])na$	$1	0	<[NV].*>	<&>
#([eiéí][^aeiouáéíóú-]*[^aeiouáéíóús-])se(an)?$	$1	0	<[NV].*>	<&>
#([eiéí][^aeiouáéíóú-]*[^aeiouáéíóún-])ne$	$1	0	<[NV].*>	<&>
#([aouáóú][^aeiouáéíóú-]*s)-san?$		$1	0	<[NV].*>	<&>
#([aouáóú][^aeiouáéíóú-]*n)-na$			$1	0	<[NV].*>	<&>
#([eiéí][^aeiouáéíóú-]*s)-se(an)?$		$1	0	<[NV].*>	<&>
#([eiéí][^aeiouáéíóú-]*n)-ne$			$1	0	<[NV].*>	<&>
##  end emphatic suffixes
#############################################################################
#   common spelling errors
# iri$			irí		2	<N.*>	<&>
uil$	úil	2
iúl$	iúil	2
(.[^aeiouáéíóú])ala$	$1ála	2
ula$	úla	2
eail$	eáil	2
([^ae])oir$	$1óir	2
([^ae])ora$	$1óra	2
eor$	eoir	2
itear$	ítear	2
iteá$	íteá	2
fea$	feá	2
#io			ío		2	<.+>	<&>		# too general (fios, sioc, etc.)
aio	aío	2
([^áéóú])ionn$	$1íonn	2
([^áéóú])ioch	$1íoch	2
([^áéóú])io([bdm])h	$1ío$2h	2
([^áéóú])io([dm])ar$	$1ío$2ar	2
([^e])oidh$	$1óidh	2
([^e])ofar$	$1ófar	2
([^eií])og$	$1óg	2
# next batch from ngramprobs.pl
u$	ú	2
#chtai$			chtaí		2	<N.*>	<&>
#ai$			aí		2	<.+>	<&>
(...)i$	$1í	2
iu	iú	2
no	nó	2
leir	léir	2
^(n?dh?)eá-	$1ea-	2
^(g?ch?)om([^haómnpr]..)	$1omh$2	2
éá	éa	2
([áéóú])í	$1i	2
íó	ío	2
^bf	bhf	2
(..)fheadh$	$1feadh	2
(..)fhadh$	$1fadh	2
# places where the second vowel is the one with the fada are rarer - most of
# them are "aí", then "eá", "iú", "uí", then noise -- all but "iú" conflict
# with a resolution above; áí is, e.g., almost always a mistake for "ái" !
í([áú])	i$1	2
# i$			í		2	<.+>	<&>  # lotsa false compounds with this
#############################################################################
#  rules with somewhat lower probability (e.g. fo-, must be after non-stnd
#  verb endings to avoid "foluíonn"  parsing as "fo+luíonn"
#^do([lnr]..)		$1		0	<[AN].*>	<&>
#^fo([lnr]..)		$1		0	<[ANV].*>	<&>
#^h?in([^bcfgmp-]..)	$1		0	<[AN].*>	<&>
# these next ones work for proper names like "NGael"; otherwise
# they're caught by general "make everything lower" catch all that follows
^BP	bP	1
^BHF	bhF	1
^DT	dT	1
^GC	gC	1
^MB	mB	1
^ND	nD	1
^NG	nG	1
^TS	tS	1
# last resort for weird caps
([A-ZÁÉÍÓÚ])	\l$1	2
#  the next rule handles a lot of stuff, including unwanted hyphens
#  after prefixes (since the prefix rules will apply after recursing)
#  and similarly unwanted hyphens before emphatic endings (CO p.126)
(.)-	$1	1
###############################################################
## DEMUTATE
## Only helps when we recurse and find a standard unmutated prefix above!
## putting caps back Dec 08 for non-words like "tSíneach", "nIónach"
## which will only be found by stripping mutation and keeping cap
#^b([Pp][^h])		$1		0	<[ANV].*>	<&>
#^bh([Ff][^h])		$1		0	<[ANV].*>	<&>
#^d([Tt][^h])		$1		0	<[ANV].*>	<&>
#^g([Cc][^h])		$1		0	<[ANV].*>	<&>
#^m([Bb][^h])		$1		0	<[ANV].*>	<&>
#^n([DdGg][^h])		$1		0	<[ANV].*>	<&>
#^t([Ss][^h])		$1		0	<[ANV].*>	<&>
#^([bcdfgmpBCDFGMP])h	$1		0	<[ANVY].*>	<&>
#^([Tt])h([^s])		$1$2		0	<[ANVY].*>	<&>   # luathscéal danger
#^([Ss])h([lnraeiouáéíóú])	$1$2	0	<[ANVY].*>	<&>
#^[bdm]'([AEIOUÁÉÍÓÚaeiouáéíóú]|[Ff]h)	$1	0	<[ANVY].*>	<&>	# d'Éamonn too
#^[nt]-([aeiouáéíóú])	$1		0	<[ANV].*>	<&>
#^[nt]([AEIOUÁÉÍÓÚ])	$1		0	<[ANV].*>	<&>