id	sid	tid	token	lemma	pos
work_lowlmiflivaeblgbyxkxrbt3nq	1	1	Transactions	transaction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	1	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	1	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	1	4	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	1	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	1	6	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	1	7	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	1	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	1	9	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	1	10	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	1	11	2013	2013	CD
work_lowlmiflivaeblgbyxkxrbt3nq	1	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	1	13	125–138	125–138	CD
work_lowlmiflivaeblgbyxkxrbt3nq	1	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	2	1	Action	Action	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	2	2	Editor	Editor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	2	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	2	4	Sharon	Sharon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	2	5	Goldwater	Goldwater	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	2	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	3	1	Submitted	submit	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	3	2	10/2012	10/2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	3	3	;	;	:
work_lowlmiflivaeblgbyxkxrbt3nq	3	4	Revised	Revised	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	3	5	3/2013	3/2013	CD
work_lowlmiflivaeblgbyxkxrbt3nq	3	6	;	;	:
work_lowlmiflivaeblgbyxkxrbt3nq	3	7	Published	publish	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	3	8	5/2013	5/2013	CD
work_lowlmiflivaeblgbyxkxrbt3nq	3	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	4	1	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	4	2	©	©	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	4	3	2013	2013	CD
work_lowlmiflivaeblgbyxkxrbt3nq	4	4	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	4	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	4	6	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	4	7	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	4	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	5	1	Modeling	model	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	5	2	Child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	5	3	Divergences	divergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	5	4	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	5	5	Adult	Adult	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	6	Grammar	Grammar	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	7	Sam	Sam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	8	Sahakian	Sahakian	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	9	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	5	11	Wisconsin	Wisconsin	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	5	13	Madison	Madison	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	14	sahakian@cs.wisc.edu	sahakian@cs.wisc.edu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	15	Benjamin	Benjamin	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	16	Snyder	Snyder	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	17	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	18	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	5	19	Wisconsin	Wisconsin	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	20	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	5	21	Madison	Madison	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	22	bsnyder@cs.wisc.edu	bsnyder@cs.wisc.edu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	23	Abstract	Abstract	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	5	24	During	during	IN
work_lowlmiflivaeblgbyxkxrbt3nq	5	25	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	5	26	course	course	NN
work_lowlmiflivaeblgbyxkxrbt3nq	5	27	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	5	28	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	5	29	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	5	30	acquisi-	acquisi-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	5	31	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	5	32	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	5	33	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	5	34	produce	produce	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	5	35	linguistic	linguistic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	5	36	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	5	37	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	5	38	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	5	39	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	5	40	conform	conform	VB
work_lowlmiflivaeblgbyxkxrbt3nq	5	41	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	5	42	adult	adult	VB
work_lowlmiflivaeblgbyxkxrbt3nq	5	43	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	5	44	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	6	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	6	2	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	6	3	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	6	5	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	6	6	introduce	introduce	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	6	7	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	6	8	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	9	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	6	11	approach	approach	VB
work_lowlmiflivaeblgbyxkxrbt3nq	6	12	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	6	13	sys-	sys-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	6	14	tematically	tematically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	6	15	modeling	model	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	6	16	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	6	17	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	18	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	6	19	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	20	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	21	divergence	divergence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	6	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	7	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	7	2	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	7	3	consists	consist	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	7	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	7	5	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	7	6	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	7	7	tences	tence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	7	8	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	7	9	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	7	10	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	7	11	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	7	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	8	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	8	2	bridge	bridge	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	8	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	8	4	gap	gap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	8	5	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	8	6	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	8	7	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	8	8	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	8	9	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	8	10	discrim-	discrim-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	8	11	inatively	inatively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	8	12	reranked	reranke	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	8	13	noisy	noisy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	8	14	channel	channel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	8	15	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	8	16	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	8	17	translates	translate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	8	18	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	8	19	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	8	20	into	into	IN
work_lowlmiflivaeblgbyxkxrbt3nq	8	21	equivalent	equivalent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	8	22	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	8	23	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	8	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	9	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	9	2	method	method	NN
work_lowlmiflivaeblgbyxkxrbt3nq	9	3	outperforms	outperform	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	9	4	MT	MT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	9	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	9	6	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	9	7	baselines	baseline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	9	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	9	9	reducing	reduce	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	9	10	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	9	11	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	9	12	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	9	13	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	9	14	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	9	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	10	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	10	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	3	allows	allow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	10	4	us	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	10	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	10	6	chart	chart	VB
work_lowlmiflivaeblgbyxkxrbt3nq	10	7	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	10	8	aspects	aspect	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	10	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	10	10	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	11	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	10	13	longitudinal	longitudinal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	10	14	stud-	stud-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	10	15	ies	ie	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	10	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	10	17	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	10	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	10	19	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	10	20	investigate	investigate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	10	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	10	22	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	23	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	10	24	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	10	25	share	share	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	10	26	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	10	27	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	10	28	developmental	developmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	10	29	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	30	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	10	31	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	32	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	10	33	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	11	1	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	11	2	Introduction	introduction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	3	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	11	5	publication	publication	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	11	8	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	11	9	Study	Study	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	11	10	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	11	11	1973	1973	CD
work_lowlmiflivaeblgbyxkxrbt3nq	11	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	11	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	11	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	11	15	existence	existence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	17	standard	standard	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	11	18	stages	stage	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	11	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	20	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	21	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	11	22	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	11	23	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	11	24	underlying	underlie	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	11	25	assumption	assumption	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	26	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	11	28	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	29	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	11	30	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	11	31	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	32	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	11	33	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	12	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	12	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	12	3	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	4	moves	move	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	12	5	towards	towards	IN
work_lowlmiflivaeblgbyxkxrbt3nq	12	6	lan-	lan-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	12	7	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	8	mastery	mastery	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	12	10	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	12	11	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	12	use	use	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	13	grows	grow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	12	14	predictably	predictably	RB
work_lowlmiflivaeblgbyxkxrbt3nq	12	15	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	12	16	include	include	VB
work_lowlmiflivaeblgbyxkxrbt3nq	12	17	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	12	18	complex	complex	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	12	19	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	12	20	structures	structure	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	12	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	12	22	even-	even-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	23	tually	tually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	12	24	converging	converge	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	12	25	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	12	26	full	full	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	12	27	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	28	usage	usage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	12	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	13	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	13	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	13	3	course	course	NN
work_lowlmiflivaeblgbyxkxrbt3nq	13	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	13	5	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	13	6	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	13	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	13	8	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	13	9	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	13	10	produce	produce	VB
work_lowlmiflivaeblgbyxkxrbt3nq	13	11	linguistic	linguistic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	13	12	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	13	13	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	13	14	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	13	15	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	13	16	conform	conform	VB
work_lowlmiflivaeblgbyxkxrbt3nq	13	17	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	13	18	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	13	19	grammatical	grammatical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	13	20	standard	standard	NN
work_lowlmiflivaeblgbyxkxrbt3nq	13	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	14	1	From	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	14	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	14	3	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	4	point	point	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	14	6	view	view	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	7	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	14	8	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	14	9	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	10	er-	er-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	14	11	rors	ror	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	14	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	14	13	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	14	14	label	label	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	15	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	14	16	implies	imply	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	14	17	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	14	18	faulty	faulty	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	14	19	production	production	NN
work_lowlmiflivaeblgbyxkxrbt3nq	14	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	15	1	Con-	Con-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	15	2	sidering	sidere	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	15	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	15	4	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	5	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	15	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	15	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	15	8	progress	progress	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	9	nature	nature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	15	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	15	12	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	13	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	14	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	15	learner	learner	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	15	17	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	15	18	divergences	divergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	15	19	could	could	MD
work_lowlmiflivaeblgbyxkxrbt3nq	15	20	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	15	21	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	15	22	de-	de-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	15	23	scribed	scribe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	15	24	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	15	25	expressions	expression	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	15	26	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	15	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	15	28	structural	structural	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	15	29	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	15	30	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	15	31	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	32	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	15	33	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	34	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	15	35	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	16	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	16	2	predictability	predictability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	16	4	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	16	5	divergences	divergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	6	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	16	7	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	16	8	observed	observe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	16	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	16	10	psychol-	psychol-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	16	11	ogists	ogist	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	16	13	linguists	linguist	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	16	15	parents	parent	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	16	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	16	17	Owens	Owens	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	16	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	16	19	2008).1	2008).1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	16	20	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	16	21	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	22	leverages	leverage	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	16	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	16	24	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	25	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	16	26	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	27	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	16	28	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	29	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	30	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	16	31	make	make	VB
work_lowlmiflivaeblgbyxkxrbt3nq	16	32	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	16	33	contributions	contribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	34	to-	to-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	16	35	wards	ward	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	16	36	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	16	37	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	38	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	16	39	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	40	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	16	41	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	17	1	First	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	17	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	17	3	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	17	4	provide	provide	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	17	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	17	6	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	17	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	17	8	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	17	9	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	17	10	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	17	11	anno-	anno-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	17	12	tated	tat	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	17	13	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	17	14	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	17	15	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	17	16	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	17	17	rephrasings	rephrasing	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	17	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	18	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	18	2	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	18	3	will	will	MD
work_lowlmiflivaeblgbyxkxrbt3nq	18	4	al-	al-	VB
work_lowlmiflivaeblgbyxkxrbt3nq	18	5	low	low	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	18	6	researchers	researcher	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	18	7	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	18	8	test	test	VB
work_lowlmiflivaeblgbyxkxrbt3nq	18	9	hypotheses	hypothesis	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	18	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	18	11	build	build	VB
work_lowlmiflivaeblgbyxkxrbt3nq	18	12	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	18	13	relating	relate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	18	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	18	15	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	18	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	18	17	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	18	18	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	18	19	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	18	20	adult	adult	VB
work_lowlmiflivaeblgbyxkxrbt3nq	18	21	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	18	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	19	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	19	2	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	19	3	contribution	contribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	19	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	19	6	probabilistic	probabilistic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	19	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	8	trained	train	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	19	9	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	19	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	19	11	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	12	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	19	13	predicts	predict	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	19	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	19	15	gram-	gram-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	19	16	matical	matical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	19	17	rephrasing	rephrasing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	18	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	19	19	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	19	20	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	19	21	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	22	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	19	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	20	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	20	2	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	20	3	assumption	assumption	NN
work_lowlmiflivaeblgbyxkxrbt3nq	20	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	20	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	20	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	20	7	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	20	8	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	20	9	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	20	10	begin	begin	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	20	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	20	12	underlying	underlie	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	20	13	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	20	14	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	20	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	20	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	20	17	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	20	18	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	20	19	stochastically	stochastically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	20	20	transformed	transform	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	20	21	into	into	IN
work_lowlmiflivaeblgbyxkxrbt3nq	20	22	observed	observed	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	20	23	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	20	24	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	20	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	21	1	Given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	21	2	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	21	3	observed	observed	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	21	4	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	5	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	6	s	s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	21	8	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	21	9	calculate	calculate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	21	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	21	11	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	21	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	21	14	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	21	15	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	16	trans-	trans-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	17	lation	lation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	18	t	t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	19	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	21	20	P(t|s	P(t|s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	21	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	21	22	∝	∝	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	23	P(s|t)P(t	P(s|t)P(t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	24	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	21	25	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	21	26	where	where	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	21	27	P(t	P(t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	28	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	21	29	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	21	30	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	21	31	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	32	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	33	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	34	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	21	35	P(s|t	P(s|t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	21	36	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	21	37	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	21	38	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	21	39	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	40	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	41	crafted	craft	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	21	42	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	21	43	capture	capture	VB
work_lowlmiflivaeblgbyxkxrbt3nq	21	44	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	45	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	46	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	21	47	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	21	48	omission	omission	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	49	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	21	50	certain	certain	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	21	51	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	52	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	21	53	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	21	54	corruptions	corruption	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	21	55	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	21	56	tense	tense	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	57	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	21	58	declension	declension	NN
work_lowlmiflivaeblgbyxkxrbt3nq	21	59	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	22	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	22	2	parame-	parame-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	3	ters	ter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	22	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	22	5	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	22	6	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	8	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	22	9	estimated	estimate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	22	10	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	22	11	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	22	12	cor-	cor-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	13	pus	pus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	22	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	22	17	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	18	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	22	19	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	20	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	22	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	22	22	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	22	23	EM	EM	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	22	24	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	22	25	handle	handle	VB
work_lowlmiflivaeblgbyxkxrbt3nq	22	26	unobserved	unobserved	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	22	27	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	22	28	alignments	alignment	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	22	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	23	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	23	2	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	23	3	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	23	4	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	23	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	6	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	23	7	produce	produce	VB
work_lowlmiflivaeblgbyxkxrbt3nq	23	8	n	n	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	23	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	23	10	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	23	11	lists	list	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	23	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	23	13	candi-	candi-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	23	14	date	date	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	15	corrections	correction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	23	16	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	23	17	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	23	18	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	23	19	reranked	reranke	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	23	20	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	23	21	long	long	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	23	22	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	23	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	24	features	feature	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	23	25	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	23	26	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	23	27	discriminative	discriminative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	23	28	frame-	frame-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	29	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	23	30	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	23	31	Collins	Collins	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	23	32	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	23	33	Roark	Roark	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	23	34	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	23	35	2004	2004	CD
work_lowlmiflivaeblgbyxkxrbt3nq	23	36	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	23	37	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	24	1	1For	1for	LS
work_lowlmiflivaeblgbyxkxrbt3nq	24	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	24	3	remainder	remainder	NN
work_lowlmiflivaeblgbyxkxrbt3nq	24	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	24	5	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	24	6	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	24	7	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	24	8	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	24	9	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	24	10	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	24	11	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	24	12	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	24	13	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	24	14	diver-	diver-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	24	15	gence	gence	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	24	16	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	24	17	interchangeably	interchangeably	RB
work_lowlmiflivaeblgbyxkxrbt3nq	24	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	25	1	125	125	CD
work_lowlmiflivaeblgbyxkxrbt3nq	25	2	One	one	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	25	3	could	could	MD
work_lowlmiflivaeblgbyxkxrbt3nq	25	4	argue	argue	VB
work_lowlmiflivaeblgbyxkxrbt3nq	25	5	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	6	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	25	7	noisy	noisy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	25	8	channel	channel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	25	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	10	mirrors	mirror	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	25	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	25	12	cognitive	cognitive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	25	13	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	16	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	17	pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	18	duction	duction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	19	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	20	appealing	appeal	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	25	21	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	22	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	25	23	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	24	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	25	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	25	26	rapidly	rapidly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	25	27	learn	learn	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	25	28	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	29	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	25	30	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	25	31	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	32	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	25	33	produce	produce	VB
work_lowlmiflivaeblgbyxkxrbt3nq	25	34	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	25	35	due	due	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	36	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	25	37	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	38	factors	factor	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	25	39	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	25	40	Bloom	bloom	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	41	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	25	42	1990	1990	CD
work_lowlmiflivaeblgbyxkxrbt3nq	25	43	;	;	:
work_lowlmiflivaeblgbyxkxrbt3nq	25	44	Ham-	Ham-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	25	45	burger	burger	NN
work_lowlmiflivaeblgbyxkxrbt3nq	25	46	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	25	47	Crain	Crain	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	25	48	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	25	49	1984	1984	CD
work_lowlmiflivaeblgbyxkxrbt3nq	25	50	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	25	51	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	26	1	That	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	26	2	being	be	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	26	3	said	say	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	26	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	26	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	26	6	pri-	pri-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	26	7	mary	mary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	26	8	goal	goal	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	26	10	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	26	11	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	12	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	26	13	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	26	14	cognitive	cognitive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	26	15	plausibility	plausibility	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	26	17	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	26	18	rather	rather	RB
work_lowlmiflivaeblgbyxkxrbt3nq	26	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	26	20	creation	creation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	21	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	26	22	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	26	23	practical	practical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	26	24	tool	tool	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	25	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	26	26	aid	aid	VB
work_lowlmiflivaeblgbyxkxrbt3nq	26	27	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	26	28	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	26	29	empirical	empirical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	26	30	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	31	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	26	32	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	33	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	26	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	27	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	2	au-	au-	CC
work_lowlmiflivaeblgbyxkxrbt3nq	27	3	tomatically	tomatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	27	4	inferring	infer	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	27	5	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	27	7	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	27	8	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	10	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	11	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	12	tences	tence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	27	14	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	27	15	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	16	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	27	17	highlight	highlight	VB
work_lowlmiflivaeblgbyxkxrbt3nq	27	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	27	19	compare	compare	VB
work_lowlmiflivaeblgbyxkxrbt3nq	27	20	devel-	devel-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	27	21	opmental	opmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	27	22	trends	trend	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	24	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	25	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	26	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	27	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	27	28	large	large	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	27	29	quantities	quantity	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	30	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	31	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	27	32	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	27	33	while	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	34	minimizing	minimize	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	27	35	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	27	36	need	need	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	37	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	27	38	hu-	hu-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	27	39	man	man	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	40	annotation	annotation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	27	41	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	28	1	Besides	besides	IN
work_lowlmiflivaeblgbyxkxrbt3nq	28	2	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	28	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	28	4	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	28	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	28	6	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	28	7	predictive	predictive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	28	8	success	success	NN
work_lowlmiflivaeblgbyxkxrbt3nq	28	9	it-	it-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	28	10	self	self	NN
work_lowlmiflivaeblgbyxkxrbt3nq	28	11	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	28	12	theoretical	theoretical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	28	13	implications	implication	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	28	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	29	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	29	2	aggregating	aggregate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	29	3	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	29	4	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	29	5	testing	testing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	29	6	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	29	7	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	29	8	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	29	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	29	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	29	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	29	12	instantiates	instantiate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	29	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	29	14	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	29	15	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	29	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	29	17	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	29	18	shared	share	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	29	19	de-	de-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	29	20	velopmental	velopmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	29	21	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	29	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	30	1	Even	even	RB
work_lowlmiflivaeblgbyxkxrbt3nq	30	2	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	30	3	adequate	adequate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	30	4	per	per	IN
work_lowlmiflivaeblgbyxkxrbt3nq	30	5	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	30	6	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	7	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	8	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	30	9	exists	exist	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	30	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	30	11	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	30	12	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	30	13	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	30	14	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	30	15	other	other	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	30	16	chil-	chil-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	30	17	dren	dren	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	18	leads	lead	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	30	19	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	30	20	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	30	21	degradation	degradation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	22	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	30	23	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	24	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	30	25	sug-	sug-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	30	26	gesting	geste	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	30	27	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	30	28	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	30	29	learned	learn	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	30	30	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	30	31	capture	capture	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	30	32	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	30	33	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	34	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	30	35	phenomena	phenomenon	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	30	36	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	30	37	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	30	38	just	just	RB
work_lowlmiflivaeblgbyxkxrbt3nq	30	39	individual	individual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	30	40	habits	habit	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	30	41	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	31	1	Besides	besides	IN
work_lowlmiflivaeblgbyxkxrbt3nq	31	2	aggregating	aggregate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	31	3	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	31	4	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	31	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	31	6	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	31	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	31	8	coarsely	coarsely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	31	9	lumps	lump	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	31	10	together	together	RB
work_lowlmiflivaeblgbyxkxrbt3nq	31	11	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	31	12	stages	stage	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	31	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	31	14	devel-	devel-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	31	15	opment	opment	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	31	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	31	17	providing	provide	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	31	18	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	31	19	frozen	frozen	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	31	20	snapshot	snapshot	NN
work_lowlmiflivaeblgbyxkxrbt3nq	31	21	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	31	22	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	31	23	gram-	gram-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	31	24	mar	mar	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	31	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	32	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	32	2	establishes	establish	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	32	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	32	4	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	32	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	32	6	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	32	7	cognitively	cognitively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	32	8	plausible	plausible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	32	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	32	10	temporally	temporally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	32	11	dynamic	dynamic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	32	12	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	32	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	33	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	33	2	compare	compare	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	33	3	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	33	4	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	5	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	6	against	against	IN
work_lowlmiflivaeblgbyxkxrbt3nq	33	7	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	33	8	baselines	baseline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	33	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	33	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	33	11	phrase	phrase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	33	13	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	33	14	Machine	Machine	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	15	Translation	Translation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	16	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	33	17	MT	MT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	18	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	33	19	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	33	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	33	22	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	33	23	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	24	designed	design	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	33	25	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	33	26	English	English	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	27	Second	Second	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	28	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	29	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	33	30	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	33	31	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	33	32	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	33	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	33	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	34	1	Relative	relative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	34	2	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	34	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	34	4	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	34	5	performing	performing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	6	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	34	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	34	9	approach	approach	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	10	achieves	achieve	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	34	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	34	12	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	34	13	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	14	decrease	decrease	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	34	16	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	17	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	18	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	34	19	rate	rate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	20	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	34	21	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	34	22	four	four	CD
work_lowlmiflivaeblgbyxkxrbt3nq	34	23	point	point	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	24	increase	increase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	25	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	34	26	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	34	27	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	34	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	35	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	35	2	analyze	analyze	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	35	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	35	4	perfor-	perfor-	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	35	5	mance	mance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	35	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	35	8	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	9	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	35	10	various	various	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	35	11	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	12	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	13	cate-	cate-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	14	gories	gorie	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	35	16	highlighting	highlight	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	35	17	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	35	18	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	19	’s	’s	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	20	strengths	strength	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	21	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	35	22	correct-	correct-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	35	23	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	35	24	be	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	35	25	drops	drop	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	26	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	35	27	morphological	morphological	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	35	28	overgeneralizations	overgeneralization	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	29	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	35	30	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	35	31	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	35	32	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	35	33	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	35	34	weaknesses	weakness	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	35	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	35	36	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	35	37	pronoun	pronoun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	35	38	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	35	39	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	35	40	drops	drop	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	35	41	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	35	42	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	36	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	36	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	36	3	assess	assess	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	36	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	36	5	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	36	6	rate	rate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	36	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	36	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	36	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	36	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	36	11	showing	show	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	36	12	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	36	13	very	very	RB
work_lowlmiflivaeblgbyxkxrbt3nq	36	14	little	little	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	36	15	annotation	annotation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	36	16	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	36	17	needed	need	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	36	18	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	36	19	achieve	achieve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	36	20	high	high	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	36	21	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	36	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	37	1	Finally	finally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	37	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	37	3	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	37	4	showcase	showcase	VB
work_lowlmiflivaeblgbyxkxrbt3nq	37	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	37	6	potential	potential	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	37	7	application	application	NN
work_lowlmiflivaeblgbyxkxrbt3nq	37	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	37	9	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	37	10	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	37	11	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	37	12	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	37	13	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	37	14	chart	chart	VB
work_lowlmiflivaeblgbyxkxrbt3nq	37	15	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	37	16	aspect	aspect	NN
work_lowlmiflivaeblgbyxkxrbt3nq	37	17	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	37	18	four	four	CD
work_lowlmiflivaeblgbyxkxrbt3nq	37	19	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	37	20	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	37	21	grammar	grammar	VB
work_lowlmiflivaeblgbyxkxrbt3nq	37	22	ac-	ac-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	37	23	quisition	quisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	37	24	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	37	25	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	37	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	38	1	While	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	38	2	generally	generally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	38	3	vindicating	vindicate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	38	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	38	5	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	38	6	thesis	thesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	38	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	38	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	38	9	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	38	10	developmental	developmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	38	11	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	38	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	38	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	38	14	results	result	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	38	15	point	point	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	38	16	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	38	17	subtleties	subtlety	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	38	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	38	19	variation	variation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	38	20	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	38	21	individ-	individ-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	38	22	uals	ual	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	38	23	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	38	24	merit	merit	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	38	25	further	further	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	38	26	investigation	investigation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	38	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	39	1	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	39	2	Background	Background	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	39	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	39	4	Related	Related	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	39	5	Work	Work	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	39	6	While	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	39	7	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	8	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	9	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	10	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	39	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	39	12	novel	novel	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	39	13	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	39	15	com-	com-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	16	putational	putational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	39	17	methods	method	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	39	18	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	39	19	frequently	frequently	RB
work_lowlmiflivaeblgbyxkxrbt3nq	39	20	used	use	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	39	21	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	39	22	study	study	VB
work_lowlmiflivaeblgbyxkxrbt3nq	39	23	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	39	24	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	25	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	39	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	40	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	40	2	computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	40	3	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	40	5	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	40	7	facilitated	facilitate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	40	8	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	40	9	TalkBank	TalkBank	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	40	10	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	40	11	MacWhinney	MacWhinney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	40	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	40	13	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	40	14	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	40	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	40	16	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	40	17	large	large	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	40	18	database	database	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	40	20	transcribed	transcribed	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	40	21	dialogues	dialogue	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	40	22	in-	in-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	40	23	cluding	clude	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	40	24	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	40	25	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	40	26	MacWhinney	MacWhinney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	40	27	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	40	28	2000	2000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	40	29	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	40	30	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	40	31	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	40	32	subsec-	subsec-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	40	33	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	34	composed	compose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	40	35	entirely	entirely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	40	36	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	40	37	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	38	conversation	conversation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	40	39	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	40	40	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	41	1	Computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	41	2	tools	tool	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	41	3	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	41	4	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	41	5	developed	develop	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	41	6	specif-	specif-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	41	7	ically	ically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	41	8	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	41	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	41	10	large	large	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	41	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	41	12	scale	scale	NN
work_lowlmiflivaeblgbyxkxrbt3nq	41	13	analysis	analysis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	41	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	41	15	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	41	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	42	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	42	2	tools	tool	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	42	3	enable	enable	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	42	4	further	further	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	5	computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	6	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	7	such	such	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	8	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	42	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	42	10	automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	11	calculation	calculation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	42	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	42	14	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	15	devel-	devel-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	16	opment	opment	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	17	metrics	metrics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	18	IPSYN	IPSYN	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	19	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	20	Sagae	Sagae	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	21	et	et	FW
work_lowlmiflivaeblgbyxkxrbt3nq	42	22	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	23	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	24	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	42	25	2005	2005	CD
work_lowlmiflivaeblgbyxkxrbt3nq	42	26	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	27	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	42	28	D	d	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	29	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	42	30	Level	level	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	31	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	32	Lu	Lu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	42	34	2009	2009	CD
work_lowlmiflivaeblgbyxkxrbt3nq	42	35	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	36	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	42	37	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	42	38	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	42	39	automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	40	formula-	formula-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	41	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	42	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	42	43	novel	novel	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	42	44	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	45	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	46	metrics	metric	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	42	47	them-	them-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	42	48	selves	self	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	42	49	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	50	Sahakian	Sahakian	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	51	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	42	52	Snyder	Snyder	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	42	53	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	42	54	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	42	55	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	42	56	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	43	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	43	2	availability	availability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	4	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	43	7	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	43	8	key	key	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	43	9	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	43	11	design	design	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	13	computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	43	14	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	43	15	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	16	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	17	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	18	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	19	Alishahi	Alishahi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	43	21	2010	2010	CD
work_lowlmiflivaeblgbyxkxrbt3nq	43	22	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	43	24	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	43	25	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	43	26	support	support	VB
work_lowlmiflivaeblgbyxkxrbt3nq	43	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	43	28	plausibility	plausibility	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	29	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	30	proposed	propose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	43	31	human	human	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	43	32	strategies	strategy	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	43	33	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	34	tasks	task	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	43	35	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	43	36	semantic	semantic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	43	37	role	role	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	38	labeling	labeling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	39	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	40	Connor	Connor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	41	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	42	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	43	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	44	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	43	45	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	43	46	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	47	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	43	48	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	49	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	43	50	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	51	Regier	Regier	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	43	52	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	43	53	2005	2005	CD
work_lowlmiflivaeblgbyxkxrbt3nq	43	54	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	43	55	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	44	1	To	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	44	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	44	3	knowledge	knowledge	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	4	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	44	5	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	44	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	44	8	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	44	9	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	10	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	44	11	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	12	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	44	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	44	15	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	44	16	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	17	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	44	18	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	44	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	45	1	Previous	previous	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	45	2	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	45	3	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	45	4	employed	employ	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	45	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	45	6	classifier	classifier	NN
work_lowlmiflivaeblgbyxkxrbt3nq	45	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	45	8	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	45	9	approach	approach	NN
work_lowlmiflivaeblgbyxkxrbt3nq	45	10	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	45	11	iden-	iden-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	45	12	tify	tify	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	45	13	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	45	14	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	45	15	indicative	indicative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	45	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	45	17	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	45	18	disorders	disorder	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	45	19	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	45	20	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	45	21	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	45	22	Morley	Morley	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	45	23	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	45	24	Prud’hommeaux	Prud’hommeaux	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	45	25	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	45	26	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	45	27	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	45	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	46	1	Automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	46	2	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	46	4	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	46	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	6	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	46	7	L2	L2	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	46	8	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	46	9	writing	writing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	10	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	46	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	46	12	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	46	13	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	14	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	46	15	computer	computer	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	16	assisted	assist	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	46	17	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	18	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	46	19	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	46	20	CALL	CALL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	46	21	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	46	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	47	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	47	2	tasks	task	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	47	3	generally	generally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	47	4	target	target	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	47	5	high	high	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	47	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	47	7	frequency	frequency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	8	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	9	categories	category	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	47	10	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	47	11	ar-	ar-	CC
work_lowlmiflivaeblgbyxkxrbt3nq	47	12	ticle	ticle	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	47	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	15	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	47	16	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	47	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	47	19	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	20	choice	choice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	47	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	48	1	Previous	previous	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	48	2	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	4	CALL	CALL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	5	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	6	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	7	includes	include	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	48	8	identify-	identify-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	9	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	10	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	11	choice	choice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	12	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	14	TOEFL	TOEFL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	15	essays	essay	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	16	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	48	17	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	18	context	context	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	19	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	20	Chodorow	Chodorow	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	48	22	Leacock	Leacock	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	48	24	2000	2000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	48	25	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	48	27	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	48	28	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	29	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	30	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	48	31	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	48	32	lattice	lattice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	33	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	48	34	PCFG	pcfg	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	35	rerank-	rerank-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	36	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	37	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	38	Lee	Lee	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	39	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	48	40	Seneff	Seneff	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	41	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	48	42	2006	2006	CD
work_lowlmiflivaeblgbyxkxrbt3nq	48	43	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	44	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	48	45	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	48	46	identifying	identify	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	48	47	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	48	48	broad	broad	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	48	49	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	50	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	51	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	52	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	53	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	54	essays	essay	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	55	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	56	examining	examine	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	48	57	linguis-	linguis-	.
work_lowlmiflivaeblgbyxkxrbt3nq	48	58	tic	tic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	48	59	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	60	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	61	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	48	62	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	48	63	sequence	sequence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	48	64	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	65	Gamon	Gamon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	48	66	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	48	67	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	48	68	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	48	69	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	49	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	49	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	49	3	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	49	4	shared	share	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	49	5	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	6	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	7	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	8	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	49	9	Dale	Dale	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	49	11	Kilgar-	Kilgar-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	12	riff	riff	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	49	14	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	49	15	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	49	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	49	17	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	49	18	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	49	19	performing	perform	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	49	20	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	21	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	49	22	Rozovskaya	Rozovskaya	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	23	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	24	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	25	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	49	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	49	27	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	49	28	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	49	29	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	49	30	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	31	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	49	32	article	article	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	49	34	punctu-	punctu-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	49	35	ation	ation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	36	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	49	37	spelling	spelling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	38	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	49	39	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	49	40	building	build	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	49	41	classifiers	classifier	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	49	42	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	49	43	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	49	44	category	category	NN
work_lowlmiflivaeblgbyxkxrbt3nq	49	45	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	50	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	50	2	line	line	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	50	4	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	50	6	grounded	ground	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	50	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	50	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	50	9	practical	practical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	50	10	application	application	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	11	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	50	12	automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	50	13	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	14	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	15	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	50	16	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	50	17	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	50	18	tool	tool	NN
work_lowlmiflivaeblgbyxkxrbt3nq	50	19	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	50	20	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	50	21	students	student	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	50	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	51	1	Statistical	statistical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	51	2	Machine	Machine	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	3	Translation	Translation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	4	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	5	SMT	SMT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	6	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	7	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	51	8	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	51	9	126	126	CD
work_lowlmiflivaeblgbyxkxrbt3nq	51	10	applied	apply	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	51	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	51	12	diverse	diverse	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	51	13	contexts	contexts	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	14	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	51	15	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	16	cor-	cor-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	51	17	rection	rection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	18	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	51	19	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	51	20	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	51	21	paraphrasing	paraphrase	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	51	22	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	23	Quirk	Quirk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	24	et	et	FW
work_lowlmiflivaeblgbyxkxrbt3nq	51	25	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	26	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	27	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	51	28	2004	2004	CD
work_lowlmiflivaeblgbyxkxrbt3nq	51	29	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	30	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	51	31	question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	32	answering	answering	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	33	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	34	Echihabi	Echihabi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	35	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	51	36	Marcu	Marcu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	37	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	51	38	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	51	39	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	40	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	51	41	prediction	prediction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	42	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	51	43	twitter	twitter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	51	44	responses	response	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	51	45	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	46	Ritter	Ritter	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	47	et	et	FW
work_lowlmiflivaeblgbyxkxrbt3nq	51	48	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	49	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	51	50	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	51	51	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	51	52	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	51	53	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	52	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	52	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	52	3	realm	realm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	52	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	52	5	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	52	6	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	52	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	52	8	SMT	SMT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	52	9	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	52	10	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	52	11	ap-	ap-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	52	12	plied	ply	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	52	13	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	52	14	identify	identify	VB
work_lowlmiflivaeblgbyxkxrbt3nq	52	15	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	52	16	correct	correct	VB
work_lowlmiflivaeblgbyxkxrbt3nq	52	17	spelling	spelling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	52	18	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	52	19	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	52	20	inter-	inter-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	52	21	net	net	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	52	22	search	search	NN
work_lowlmiflivaeblgbyxkxrbt3nq	52	23	queries	query	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	52	24	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	52	25	Sun	Sun	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	52	26	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	52	27	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	52	28	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	52	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	52	30	2010	2010	CD
work_lowlmiflivaeblgbyxkxrbt3nq	52	31	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	52	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	53	1	Within	within	IN
work_lowlmiflivaeblgbyxkxrbt3nq	53	2	CALL	CALL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	53	4	Park	Park	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	53	6	Levy	Levy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	7	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	53	8	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	53	9	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	53	10	took	take	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	53	11	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	53	12	unsupervised	unsupervised	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	53	13	SMT	SMT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	14	approach	approach	NN
work_lowlmiflivaeblgbyxkxrbt3nq	53	15	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	53	16	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	17	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	53	18	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	53	19	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	53	20	Weighted	Weighted	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	21	Fi-	Fi-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	22	nite	nite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	53	23	State	State	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	24	Transducers	Transducers	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	53	25	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	53	26	FSTs	FSTs	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	53	27	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	53	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	54	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	54	2	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	54	3	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	54	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	54	5	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	54	6	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	54	7	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	54	8	inspired	inspire	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	54	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	54	10	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	54	11	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	54	12	Park	Park	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	54	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	54	14	Levy	Levy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	54	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	54	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	54	17	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	54	18	Section	section	NN
work_lowlmiflivaeblgbyxkxrbt3nq	54	19	6	6	CD
work_lowlmiflivaeblgbyxkxrbt3nq	54	20	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	54	21	detail	detail	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	54	22	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	54	23	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	54	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	54	25	approaches	approach	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	54	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	55	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	55	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	55	3	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	55	4	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	55	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	55	6	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	55	7	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	55	8	base-	base-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	55	9	line	line	NN
work_lowlmiflivaeblgbyxkxrbt3nq	55	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	56	1	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	56	2	Data	Data	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	56	3	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	56	4	train	train	VB
work_lowlmiflivaeblgbyxkxrbt3nq	56	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	56	6	evaluate	evaluate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	56	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	56	8	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	9	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	56	11	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	56	12	first	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	56	13	collected	collect	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	56	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	56	15	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	56	17	1,000	1,000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	56	18	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	56	19	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	20	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	56	21	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	22	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	56	23	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	56	24	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	56	25	American	american	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	56	26	English	English	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	56	27	portion	portion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	28	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	56	29	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	56	30	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	56	31	database	database	NN
work_lowlmiflivaeblgbyxkxrbt3nq	56	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	57	1	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	57	2	encourage	encourage	VB
work_lowlmiflivaeblgbyxkxrbt3nq	57	3	diversity	diversity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	57	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	57	6	grammatical	grammatical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	57	7	divergences	divergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	57	8	captured	capture	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	57	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	57	11	corpus	corpus	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	57	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	57	13	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	57	14	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	57	15	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	57	16	drawn	draw	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	57	17	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	18	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	57	19	large	large	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	57	20	pool	pool	NN
work_lowlmiflivaeblgbyxkxrbt3nq	57	21	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	22	studies	study	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	57	23	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	57	24	see	see	VB
work_lowlmiflivaeblgbyxkxrbt3nq	57	25	bibliography	bibliography	NN
work_lowlmiflivaeblgbyxkxrbt3nq	57	26	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	57	28	full	full	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	57	29	list	list	NN
work_lowlmiflivaeblgbyxkxrbt3nq	57	30	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	57	31	citations	citation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	57	32	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	57	33	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	58	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	58	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	58	3	annotation	annotation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	4	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	58	6	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	7	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	8	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	9	tences	tence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	58	10	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	58	11	randomly	randomly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	58	12	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	58	13	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	58	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	58	15	pool	pool	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	58	17	classified	classify	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	58	18	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	58	19	hand	hand	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	20	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	58	21	either	either	CC
work_lowlmiflivaeblgbyxkxrbt3nq	58	22	grammatically	grammatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	58	23	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	58	24	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	58	25	divergent	divergent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	58	26	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	58	27	unclassifiable	unclassifiable	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	58	28	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	58	29	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	58	30	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	58	31	was	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	58	32	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	58	33	possi-	possi-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	58	34	ble	ble	VB
work_lowlmiflivaeblgbyxkxrbt3nq	58	35	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	58	36	tell	tell	VB
work_lowlmiflivaeblgbyxkxrbt3nq	58	37	what	what	WP
work_lowlmiflivaeblgbyxkxrbt3nq	58	38	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	58	39	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	58	40	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	58	41	trying	try	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	58	42	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	58	43	say	say	VB
work_lowlmiflivaeblgbyxkxrbt3nq	58	44	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	58	45	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	59	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	59	2	con-	con-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	59	3	tinued	tinue	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	59	4	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	59	5	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	59	6	until	until	IN
work_lowlmiflivaeblgbyxkxrbt3nq	59	7	1,000	1,000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	59	8	divergent	divergent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	59	9	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	59	10	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	59	11	found	find	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	59	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	60	1	Along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	60	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	60	3	way	way	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	4	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	60	5	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	60	6	encountered	encounter	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	60	7	5,197	5,197	CD
work_lowlmiflivaeblgbyxkxrbt3nq	60	8	grammatically	grammatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	60	9	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	60	10	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	60	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	60	12	909	909	CD
work_lowlmiflivaeblgbyxkxrbt3nq	60	13	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	60	14	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	60	15	unclassifiable.2	unclassifiable.2	XX
work_lowlmiflivaeblgbyxkxrbt3nq	60	16	Because	because	IN
work_lowlmiflivaeblgbyxkxrbt3nq	60	17	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	60	18	includes	include	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	60	19	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	20	samples	sample	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	60	21	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	60	22	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	60	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	60	24	diverse	diverse	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	60	25	age	age	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	60	27	back-	back-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	28	ground	ground	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	29	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	60	30	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	31	ability	ability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	32	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	60	33	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	60	34	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	35	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	60	36	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	60	37	capture	capture	VB
work_lowlmiflivaeblgbyxkxrbt3nq	60	38	any	any	DT
work_lowlmiflivaeblgbyxkxrbt3nq	60	39	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	60	40	stage	stage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	41	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	60	42	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	43	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	60	44	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	61	1	Instead	instead	RB
work_lowlmiflivaeblgbyxkxrbt3nq	61	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	61	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	61	4	corpus	corpus	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	61	5	represents	represent	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	61	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	61	7	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	61	8	snapshot	snapshot	NN
work_lowlmiflivaeblgbyxkxrbt3nq	61	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	61	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	61	11	learner	learner	NN
work_lowlmiflivaeblgbyxkxrbt3nq	61	12	who	who	WP
work_lowlmiflivaeblgbyxkxrbt3nq	61	13	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	61	14	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	61	15	yet	yet	RB
work_lowlmiflivaeblgbyxkxrbt3nq	61	16	mastered	master	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	61	17	English	English	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	61	18	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	61	19	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	61	20	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	61	21	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	61	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	62	1	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	62	2	provide	provide	VB
work_lowlmiflivaeblgbyxkxrbt3nq	62	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	62	4	grammatically	grammatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	62	5	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	62	6	counterpart	counterpart	NN
work_lowlmiflivaeblgbyxkxrbt3nq	62	7	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	62	8	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	62	9	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	62	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	62	11	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	62	12	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	62	13	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	62	14	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	62	15	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	62	16	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	62	17	workers	worker	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	62	18	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	62	19	Amazon	Amazon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	62	20	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	62	21	Mechanical	Mechanical	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	62	22	Turk	Turk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	62	23	web	web	NN
work_lowlmiflivaeblgbyxkxrbt3nq	62	24	ser-	ser-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	62	25	vice	vice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	62	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	63	1	Given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	63	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	63	3	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	4	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	63	6	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	63	7	surrounding	surround	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	63	8	conversational	conversational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	63	9	context	context	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	63	11	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	63	12	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	63	13	instructed	instruct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	63	14	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	63	15	translate	translate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	63	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	63	17	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	18	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	19	into	into	IN
work_lowlmiflivaeblgbyxkxrbt3nq	63	20	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	21	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	63	22	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	63	23	En-	en-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	63	24	glish	glish	NN
work_lowlmiflivaeblgbyxkxrbt3nq	63	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	64	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	64	2	limited	limit	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	64	3	eligible	eligible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	4	workers	worker	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	64	5	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	64	6	native	native	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	7	English	english	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	8	2These	2these	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	9	hand	hand	NN
work_lowlmiflivaeblgbyxkxrbt3nq	64	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	64	11	classified	classify	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	64	12	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	64	13	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	64	14	available	available	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	15	online	online	RB
work_lowlmiflivaeblgbyxkxrbt3nq	64	16	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	64	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	64	18	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	64	19	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	64	20	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	64	21	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	64	22	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	64	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	65	1	Error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	65	2	Type	Type	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	65	3	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	65	4	Utterance	Utterance	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	65	5	Insertion	Insertion	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	65	6	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	65	7	did	do	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	65	8	locked	lock	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	65	9	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	65	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	66	1	Inflection	Inflection	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	66	2	More	More	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	66	3	cookie	cookie	NN
work_lowlmiflivaeblgbyxkxrbt3nq	66	4	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	67	1	Deletion	deletion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	67	2	That	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	67	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	67	4	how	how	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	67	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	68	1	Lemma	Lemma	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	68	2	Choice	Choice	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	68	3	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	68	4	got	get	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	68	5	grain	grain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	68	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	69	1	Overgeneralization	overgeneralization	NN
work_lowlmiflivaeblgbyxkxrbt3nq	69	2	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	69	3	drawed	draw	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	69	4	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	69	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	70	1	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	70	2	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	70	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	70	4	Examples	example	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	70	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	70	6	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	70	7	types	type	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	70	8	captured	capture	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	70	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	70	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	70	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	70	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	71	1	speakers	speaker	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	71	2	residing	reside	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	71	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	71	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	71	5	US	US	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	71	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	72	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	72	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	72	3	required	require	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	72	4	anno-	anno-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	72	5	tators	tator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	72	6	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	72	7	follow	follow	VB
work_lowlmiflivaeblgbyxkxrbt3nq	72	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	72	9	brief	brief	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	72	10	tutorial	tutorial	NN
work_lowlmiflivaeblgbyxkxrbt3nq	72	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	72	12	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	72	13	they	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	72	14	prac-	prac-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	72	15	tice	tice	RB
work_lowlmiflivaeblgbyxkxrbt3nq	72	16	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	72	17	sample	sample	NN
work_lowlmiflivaeblgbyxkxrbt3nq	72	18	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	72	19	according	accord	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	72	20	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	72	21	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	72	22	guidelines	guideline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	72	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	73	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	73	2	guidelines	guideline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	73	3	instructed	instruct	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	73	4	workers	worker	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	73	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	73	6	minimally	minimally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	73	7	alter	alter	VB
work_lowlmiflivaeblgbyxkxrbt3nq	73	8	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	73	9	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	73	10	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	73	11	grammatically	grammatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	73	12	con-	con-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	73	13	sistent	sistent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	73	14	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	73	15	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	73	16	conversation	conversation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	73	17	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	73	18	written	write	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	73	19	letter	letter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	73	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	73	21	without	without	IN
work_lowlmiflivaeblgbyxkxrbt3nq	73	22	altering	alter	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	73	23	underlying	underlie	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	73	24	meaning	meaning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	73	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	74	1	Annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	74	2	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	74	3	eval-	eval-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	74	4	uated	uated	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	74	5	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	74	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	74	7	worker	worker	NN
work_lowlmiflivaeblgbyxkxrbt3nq	74	8	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	74	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	74	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	74	11	worker	worker	NN
work_lowlmiflivaeblgbyxkxrbt3nq	74	12	basis	basis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	74	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	74	14	rejected	reject	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	74	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	74	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	74	17	rare	rare	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	74	18	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	74	19	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	74	20	they	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	74	21	ignored	ignore	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	74	22	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	74	23	guidelines	guideline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	74	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	75	1	Ac-	Ac-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	75	2	cepted	cepte	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	75	3	workers	worker	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	75	4	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	75	5	paid	pay	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	75	6	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	75	7	cents	cent	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	75	8	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	75	9	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	75	10	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	75	11	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	75	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	75	13	5	5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	75	14	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	75	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	76	1	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	76	2	achieve	achieve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	76	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	76	4	consistent	consistent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	76	5	judgment	judgment	NN
work_lowlmiflivaeblgbyxkxrbt3nq	76	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	76	7	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	76	8	posted	post	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	76	9	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	76	10	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	76	11	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	76	12	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	76	13	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	76	14	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	76	15	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	76	16	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	76	17	different	different	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	76	18	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	76	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	77	1	Once	once	IN
work_lowlmiflivaeblgbyxkxrbt3nq	77	2	multiple	multiple	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	77	3	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	77	4	translations	translation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	77	5	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	77	6	ob-	ob-	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	77	7	tained	taine	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	77	8	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	77	9	selected	select	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	77	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	77	11	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	77	12	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	77	13	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	77	14	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	77	15	plu-	plu-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	77	16	rality	rality	NN
work_lowlmiflivaeblgbyxkxrbt3nq	77	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	77	18	arbitrating	arbitrate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	77	19	ties	tie	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	77	20	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	77	21	necessary	necessary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	77	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	78	1	There	there	EX
work_lowlmiflivaeblgbyxkxrbt3nq	78	2	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	78	3	sev-	sev-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	78	4	eral	eral	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	78	5	cases	case	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	78	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	78	7	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	78	8	corrections	correction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	78	9	obtained	obtain	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	78	10	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	78	11	plurality	plurality	NN
work_lowlmiflivaeblgbyxkxrbt3nq	78	12	decision	decision	NN
work_lowlmiflivaeblgbyxkxrbt3nq	78	13	did	do	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	78	14	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	78	15	perfectly	perfectly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	78	16	follow	follow	VB
work_lowlmiflivaeblgbyxkxrbt3nq	78	17	instructions	instruction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	78	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	79	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	79	2	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	79	3	manually	manually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	79	4	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	79	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	80	1	Both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	80	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	80	3	raw	raw	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	80	4	translations	translation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	5	provided	provide	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	80	6	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	7	individual	individual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	80	8	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	9	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	80	10	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	80	11	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	80	13	cu-	cu-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	14	rated	rate	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	80	15	final	final	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	80	16	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	17	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	18	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	80	19	provided	provide	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	80	20	online	online	RB
work_lowlmiflivaeblgbyxkxrbt3nq	80	21	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	22	part	part	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	80	25	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	26	set.3	set.3	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	27	Resulting	result	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	80	28	pairs	pair	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	29	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	30	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	80	31	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	32	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	33	tences	tence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	34	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	80	35	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	80	36	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	37	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	80	38	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	80	39	corrections	correction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	40	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	80	41	split	split	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	80	42	into	into	IN
work_lowlmiflivaeblgbyxkxrbt3nq	80	43	73	73	CD
work_lowlmiflivaeblgbyxkxrbt3nq	80	44	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	45	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	46	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	80	47	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	80	48	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	49	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	50	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	80	51	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	80	52	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	53	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	54	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	80	55	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	80	56	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	80	57	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	80	58	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	80	59	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	80	60	build	build	VB
work_lowlmiflivaeblgbyxkxrbt3nq	80	61	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	80	62	tune	tune	VB
work_lowlmiflivaeblgbyxkxrbt3nq	80	63	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	80	64	evaluate	evaluate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	80	65	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	80	66	gram-	gram-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	67	mar	mar	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	80	68	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	69	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	80	70	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	81	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	81	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	81	3	final	final	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	81	4	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	81	5	phase	phase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	81	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	81	7	devel-	devel-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	81	8	opment	opment	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	81	9	data	data	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	81	10	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	81	11	included	include	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	81	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	81	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	81	14	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	81	15	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	81	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	82	1	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	82	2	Model	Model	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	82	3	According	accord	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	82	4	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	82	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	82	6	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	82	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	82	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	82	9	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	82	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	82	11	like	like	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	82	12	utter-	utter-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	82	13	ances	ance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	82	14	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	82	15	formed	form	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	82	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	82	17	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	82	18	transformed	transform	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	82	19	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	82	20	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	82	21	noisy	noisy	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	82	22	channel	channel	NN
work_lowlmiflivaeblgbyxkxrbt3nq	82	23	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	82	24	become	become	VB
work_lowlmiflivaeblgbyxkxrbt3nq	82	25	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	82	26	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	82	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	83	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	83	2	structure	structure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	83	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	83	4	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	83	5	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	83	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	83	7	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	83	8	tailored	tailor	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	83	9	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	83	10	match	match	VB
work_lowlmiflivaeblgbyxkxrbt3nq	83	11	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	83	12	observa-	observa-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	83	13	tions	tion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	83	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	83	15	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	83	16	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	83	17	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	83	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	84	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	84	2	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	84	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	84	4	func-	func-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	84	5	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	6	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	7	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	84	9	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	10	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	11	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	84	13	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	84	15	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	16	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	17	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	84	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	84	19	inflectional	inflectional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	84	20	changes	change	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	21	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	84	22	con-	con-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	84	23	tent	tent	NN
work_lowlmiflivaeblgbyxkxrbt3nq	84	24	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	84	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	85	1	Examples	example	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	85	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	85	3	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	85	4	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	85	5	type	type	NN
work_lowlmiflivaeblgbyxkxrbt3nq	85	6	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	85	7	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	85	8	3Data	3data	CD
work_lowlmiflivaeblgbyxkxrbt3nq	85	9	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	85	10	available	available	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	85	11	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	85	12	http://pages.cs.wisc.edu/~bsnyder	http://pages.cs.wisc.edu/~bsnyder	NN
work_lowlmiflivaeblgbyxkxrbt3nq	85	13	127	127	CD
work_lowlmiflivaeblgbyxkxrbt3nq	85	14	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	85	15	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	85	16	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	85	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	86	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	86	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	86	3	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	86	4	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	86	5	allow	allow	VB
work_lowlmiflivaeblgbyxkxrbt3nq	86	6	reorderings	reordering	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	86	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	86	8	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	86	9	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	86	10	thus	thus	RB
work_lowlmiflivaeblgbyxkxrbt3nq	86	11	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	86	12	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	86	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	86	14	terms	term	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	86	15	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	86	16	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	86	17	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	86	18	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	86	19	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	86	20	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	86	21	stochastic	stochastic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	86	22	transformations	transformation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	86	23	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	86	24	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	86	25	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	86	26	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	86	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	87	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	87	2	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	87	3	10	10	CD
work_lowlmiflivaeblgbyxkxrbt3nq	87	4	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	87	5	classes	class	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	6	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	87	7	parameterize	parameterize	VB
work_lowlmiflivaeblgbyxkxrbt3nq	87	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	87	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	87	10	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	87	11	pronouns	pronoun	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	13	negators	negator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	15	wh	wh	NN
work_lowlmiflivaeblgbyxkxrbt3nq	87	16	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	87	17	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	19	conjunc-	conjunc-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	87	20	tions	tion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	22	prepositions	preposition	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	24	determiners	determiner	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	25	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	26	modal	modal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	87	27	verbs	verbs	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	87	28	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	29	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	87	30	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	87	31	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	87	32	verbs	verbs	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	87	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	34	other	other	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	87	35	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	87	36	verbs	verb	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	37	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	87	38	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	87	39	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	87	40	content	content	NN
work_lowlmiflivaeblgbyxkxrbt3nq	87	41	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	87	42	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	88	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	88	2	list	list	NN
work_lowlmiflivaeblgbyxkxrbt3nq	88	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	88	4	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	88	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	88	6	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	88	7	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	88	8	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	88	9	provided	provide	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	88	10	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	88	11	part	part	NN
work_lowlmiflivaeblgbyxkxrbt3nq	88	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	88	13	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	88	14	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	88	15	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	88	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	89	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	2	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	3	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	4	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	5	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	6	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	89	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	89	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	10	generates	generate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	89	11	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	12	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	13	w′	w′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	89	14	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	15	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	16	hierarchi-	hierarchi-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	89	17	cal	cal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	89	18	series	series	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	20	draws	draw	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	89	21	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	22	multinomial	multinomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	89	23	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	89	24	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	89	25	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	89	26	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	28	original	original	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	89	29	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	30	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	89	31	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	89	32	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	89	33	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	34	c.	c.	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	35	All	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	36	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	89	37	receive	receive	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	89	38	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	39	asymmetric	asymmetric	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	89	40	Dirichlet	Dirichlet	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	89	41	prior	prior	RB
work_lowlmiflivaeblgbyxkxrbt3nq	89	42	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	89	43	favors	favor	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	89	44	retention	retention	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	45	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	89	46	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	89	47	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	48	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	89	49	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	90	1	With	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	90	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	90	3	sole	sole	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	90	4	exception	exception	NN
work_lowlmiflivaeblgbyxkxrbt3nq	90	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	90	6	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	90	7	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	90	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	90	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	90	10	distribu-	distribu-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	90	11	tions	tion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	90	12	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	90	13	parameterized	parameterize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	90	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	90	15	learned	learn	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	90	16	during	during	IN
work_lowlmiflivaeblgbyxkxrbt3nq	90	17	train-	train-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	90	18	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	90	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	91	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	91	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	91	3	consists	consist	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	91	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	91	5	217	217	CD
work_lowlmiflivaeblgbyxkxrbt3nq	91	6	multinomial	multinomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	91	7	distri-	distri-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	91	8	butions	bution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	91	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	91	10	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	91	11	6,718	6,718	CD
work_lowlmiflivaeblgbyxkxrbt3nq	91	12	free	free	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	91	13	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	91	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	92	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	92	2	precise	precise	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	92	3	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	4	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	92	5	parameterization	parameterization	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	92	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	92	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	9	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	92	10	handcrafted	handcraft	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	92	11	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	92	12	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	13	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	92	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	92	15	de-	de-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	16	velopment	velopment	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	92	17	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	92	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	92	19	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	92	20	trial	trial	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	92	22	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	92	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	93	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	93	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	93	3	con-	con-	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	93	4	sidered	sidere	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	93	5	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	93	6	fine	fine	RB
work_lowlmiflivaeblgbyxkxrbt3nq	93	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	93	8	grained	grain	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	93	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	93	10	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	93	11	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	93	12	i.e.	i.e.	FW
work_lowlmiflivaeblgbyxkxrbt3nq	94	1	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	94	2	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	3	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	94	4	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	94	5	non	non	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	94	6	-	-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	94	7	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	94	8	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	94	10	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	11	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	12	pair	pair	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	13	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	94	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	94	15	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	94	16	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	94	17	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	94	18	coarser	coarser	NN
work_lowlmiflivaeblgbyxkxrbt3nq	94	19	parameterizations	parameterization	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	94	20	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	94	21	i.e.	i.e.	FW
work_lowlmiflivaeblgbyxkxrbt3nq	95	1	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	95	2	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	95	3	shared	share	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	95	4	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	95	5	denoting	denote	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	95	6	any	any	DT
work_lowlmiflivaeblgbyxkxrbt3nq	95	7	inflection	inflection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	95	8	change	change	NN
work_lowlmiflivaeblgbyxkxrbt3nq	95	9	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	95	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	96	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	96	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	96	3	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	96	4	describe	describe	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	96	5	here	here	RB
work_lowlmiflivaeblgbyxkxrbt3nq	96	6	seemed	seem	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	96	7	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	96	8	achieve	achieve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	96	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	96	10	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	96	11	balance	balance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	96	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	96	13	specificity	specificity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	96	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	96	15	general-	general-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	96	16	ization	ization	NN
work_lowlmiflivaeblgbyxkxrbt3nq	96	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	97	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	97	2	now	now	RB
work_lowlmiflivaeblgbyxkxrbt3nq	97	3	present	present	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	97	4	pseudocode	pseudocode	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	5	describing	describe	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	97	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	97	7	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	9	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	97	10	operation	operation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	11	upon	upon	IN
work_lowlmiflivaeblgbyxkxrbt3nq	97	12	processing	process	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	97	13	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	97	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	97	16	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	97	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	97	18	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	97	19	brief	brief	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	97	20	description	description	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	21	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	97	22	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	97	23	step	step	NN
work_lowlmiflivaeblgbyxkxrbt3nq	97	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	98	1	Action	action	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	2	selection	selection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	3	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	98	4	lines	line	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	98	5	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	98	6	-	-	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	98	7	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	98	8	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	98	9	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	98	10	On	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	98	11	reading	read	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	98	12	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	98	13	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	98	16	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	98	17	action	action	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	18	category	category	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	19	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	98	20	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	98	21	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	98	22	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	98	23	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	98	24	prob-	prob-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	98	25	ability	ability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	26	distribution	distribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	27	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	98	28	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	98	29	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	98	30	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	31	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	32	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	98	33	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	98	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	99	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	99	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	99	3	allows	allow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	99	4	up	up	IN
work_lowlmiflivaeblgbyxkxrbt3nq	99	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	99	6	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	99	7	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	99	8	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	99	9	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	99	10	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	99	11	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	99	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	99	13	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	99	14	row	row	NN
work_lowlmiflivaeblgbyxkxrbt3nq	99	15	before	before	IN
work_lowlmiflivaeblgbyxkxrbt3nq	99	16	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	99	17	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	99	18	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	99	19	re-	re-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	99	20	quired	quire	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	99	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	100	1	Lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	100	2	content	content	NN
work_lowlmiflivaeblgbyxkxrbt3nq	100	3	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	100	4	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	100	5	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	100	6	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	100	7	deleted	delete	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	100	8	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	100	9	inserted	insert	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	100	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	100	11	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	100	12	swapped	swap	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	100	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	101	1	Insert	Insert	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	101	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	101	3	Delete	Delete	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	101	4	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	101	5	lines	line	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	101	6	8	8	CD
work_lowlmiflivaeblgbyxkxrbt3nq	101	7	-	-	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	101	8	15	15	CD
work_lowlmiflivaeblgbyxkxrbt3nq	101	9	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	101	10	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	101	11	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	101	12	deletion	deletion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	101	13	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	101	14	requires	require	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	101	15	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	101	16	decision	decision	NN
work_lowlmiflivaeblgbyxkxrbt3nq	101	17	after	after	IN
work_lowlmiflivaeblgbyxkxrbt3nq	101	18	action	action	NN
work_lowlmiflivaeblgbyxkxrbt3nq	101	19	selection	selection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	101	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	102	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	102	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	102	3	insertion	insertion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	4	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	102	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	102	7	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	102	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	102	10	inserted	insert	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	102	11	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	102	13	c′	c′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	102	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	102	15	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	102	16	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	102	17	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	102	18	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	102	19	cPREV	cPREV	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	102	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	102	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	102	22	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	102	24	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	102	25	previous	previous	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	102	26	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	27	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	102	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	103	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	103	2	precise	precise	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	103	3	identity	identity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	103	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	103	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	103	6	inserted	insert	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	103	7	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	103	8	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	103	9	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	103	10	drawn	draw	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	103	11	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	103	12	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	103	13	uniform	uniform	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	103	14	distribution	distribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	103	15	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	103	16	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	103	17	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	103	18	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	103	19	c′.	c′.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	104	1	It	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	104	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	104	3	important	important	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	104	4	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	104	5	note	note	VB
work_lowlmiflivaeblgbyxkxrbt3nq	104	6	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	104	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	104	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	104	9	insertion	insertion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	10	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	104	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	104	13	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	15	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	104	16	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	104	17	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	104	18	iteration	iteration	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	19	will	will	MD
work_lowlmiflivaeblgbyxkxrbt3nq	104	20	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	104	21	re	re	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	104	22	-	-	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	104	23	processed	process	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	104	24	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	104	25	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	104	26	next	next	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	104	27	iteration	iteration	NN
work_lowlmiflivaeblgbyxkxrbt3nq	104	28	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	104	29	lines	line	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	104	30	33	33	CD
work_lowlmiflivaeblgbyxkxrbt3nq	104	31	-	-	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	104	32	35	35	CD
work_lowlmiflivaeblgbyxkxrbt3nq	104	33	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	104	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	105	1	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	2	←	←	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	3	0	0	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	105	4	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	5	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	6	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	7	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	8	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	9	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	11	inflection	inflection	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	12	f	f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	14	lemma	lemma	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	15	`	`	''
work_lowlmiflivaeblgbyxkxrbt3nq	105	16	do	do	VB
work_lowlmiflivaeblgbyxkxrbt3nq	105	17	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	18	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	19	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	20	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	21	=	=	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	105	22	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	23	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	24	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	25	←	←	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	26	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	27	else	else	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	28	6	6	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	29	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	30	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	31	∼{insert	∼{insert	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	32	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	33	delete	delete	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	34	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	35	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	36	}	}	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	37	|	|	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	38	c	c	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	39	end	end	VB
work_lowlmiflivaeblgbyxkxrbt3nq	105	40	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	41	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	42	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	43	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	105	44	delete	delete	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	45	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	46	9	9	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	47	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	48	insdel++	insdel++	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	49	c′	c′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	50	←	←	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	51	�	�	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	52	w′	w′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	53	←	←	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	54	�	�	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	55	12	12	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	56	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	57	else	else	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	58	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	59	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	60	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	105	61	insert	insert	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	62	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	63	insdel++	insdel++	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	64	c′	c′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	65	∼	∼	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	66	classes	class	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	67	|	|	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	68	cPREV	cPREV	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	105	69	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	70	insert	insert	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	71	15	15	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	72	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	73	w′	w′	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	74	∼	∼	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	75	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	76	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	77	c′	c′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	78	|	|	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	79	insert	insert	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	80	else	else	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	81	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	82	←	←	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	83	0	0	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	105	84	18	18	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	85	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	86	c′	c′	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	105	87	←	←	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	88	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	89	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	90	c	c	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	91	∈	∈	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	92	uninflected	uninflected	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	105	93	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	105	94	classes	class	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	95	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	96	w′	w′	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	105	97	∼	∼	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	98	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	99	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	100	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	101	|	|	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	102	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	103	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	104	swap	swap	VB
work_lowlmiflivaeblgbyxkxrbt3nq	105	105	21	21	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	106	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	107	else	else	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	108	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	109	c	c	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	110	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	105	111	aux	aux	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	112	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	113	`	`	''
work_lowlmiflivaeblgbyxkxrbt3nq	105	114	′	′	ADD
work_lowlmiflivaeblgbyxkxrbt3nq	105	115	∼	∼	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	116	aux	aux	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	117	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	105	118	lemmas	lemmas	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	119	|	|	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	120	`	`	''
work_lowlmiflivaeblgbyxkxrbt3nq	105	121	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	122	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	123	f	f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	124	′	′	ADD
work_lowlmiflivaeblgbyxkxrbt3nq	105	125	∼	∼	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	126	inflections	inflection	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	127	|	|	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	128	f	f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	129	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	130	swap	swap	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	131	24	24	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	132	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	133	w′	w′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	134	←	←	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	135	COMBINE(`′,f	COMBINE(`′,f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	136	′	′	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	137	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	138	else	else	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	139	f	f	XX
work_lowlmiflivaeblgbyxkxrbt3nq	105	140	′	′	ADD
work_lowlmiflivaeblgbyxkxrbt3nq	105	141	∼	∼	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	142	inflections	inflection	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	143	|	|	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	144	f	f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	145	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	146	swap	swap	VB
work_lowlmiflivaeblgbyxkxrbt3nq	105	147	27	27	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	148	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	149	w′	w′	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	150	←	←	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	151	COMBINE(`,f	COMBINE(`,f	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	152	′	′	LS
work_lowlmiflivaeblgbyxkxrbt3nq	105	153	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	154	end	end	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	105	155	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	156	end	end	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	157	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	158	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	159	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	160	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	161	w′	w′	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	162	∈	∈	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	163	irregular	irregular	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	164	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	165	w′	w′	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	105	166	∼	∼	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	167	OVERGEN(w′)∪{w′	OVERGEN(w′)∪{w′	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	168	}	}	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	169	end	end	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	170	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	171	33	33	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	172	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	173	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	174	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	175	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	105	176	insert	insert	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	177	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	105	178	goto	goto	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	179	line	line	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	180	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	181	end	end	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	182	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	183	36	36	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	184	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	185	end	end	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	186	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	187	Swap	Swap	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	105	188	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	189	lines	line	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	105	190	16	16	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	191	-	-	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	105	192	29	29	CD
work_lowlmiflivaeblgbyxkxrbt3nq	105	193	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	105	194	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	105	195	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	196	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	197	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	198	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	199	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	105	200	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	201	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	202	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	203	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	105	204	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	205	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	105	206	substituted	substitute	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	105	207	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	208	another	another	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	209	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	210	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	105	211	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	105	212	same	same	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	105	213	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	105	214	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	106	1	Depending	depend	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	106	2	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	106	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	106	4	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	106	5	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	106	6	’s	’s	NN
work_lowlmiflivaeblgbyxkxrbt3nq	106	7	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	106	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	106	9	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	106	10	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	106	11	handled	handle	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	106	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	106	13	slightly	slightly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	106	14	different	different	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	106	15	ways	way	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	106	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	107	1	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	3	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	107	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	6	modal	modal	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	8	conjunction	conjunction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	10	determiner	determiner	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	12	preposi-	preposi-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	13	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	15	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	107	16	wh-	wh-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	17	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	107	18	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	19	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	107	20	negative	negative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	22	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	107	23	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	107	24	considered	consider	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	107	25	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	107	26	unin-	unin-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	27	128	128	CD
work_lowlmiflivaeblgbyxkxrbt3nq	107	28	flected	flected	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	107	30	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	107	31	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	32	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	33	cases	case	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	107	34	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	35	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	36	new	new	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	37	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	38	w′	w′	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	107	39	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	107	40	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	107	41	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	42	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	43	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	107	44	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	45	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	46	c	c	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	47	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	48	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	107	49	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	50	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	51	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	52	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	53	w.	w.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	107	54	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	55	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	107	56	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	107	57	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	58	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	59	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	60	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	107	61	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	107	62	swap	swap	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	63	procedure	procedure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	64	con-	con-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	107	65	sists	sist	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	107	66	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	107	67	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	107	68	parallel	parallel	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	107	69	steps	step	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	107	70	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	108	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	2	lemma	lemma	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	108	4	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	108	5	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	6	possible	possible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	108	7	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	108	8	lemmas	lemma	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	108	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	108	10	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	108	11	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	13	lemma	lemma	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	16	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	17	word.4	word.4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	108	18	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	20	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	108	21	step	step	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	22	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	108	23	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	24	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	25	inflection	inflection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	26	type	type	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	27	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	108	28	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	108	29	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	30	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	31	distribution	distribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	32	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	108	33	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	108	34	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	108	35	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	36	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	37	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	108	38	inflection	inflection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	108	39	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	109	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	109	2	precise	precise	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	109	3	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	109	4	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	109	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	109	6	fully	fully	RB
work_lowlmiflivaeblgbyxkxrbt3nq	109	7	specified	specify	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	109	8	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	109	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	109	10	choice	choice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	109	11	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	109	12	lemma	lemma	NN
work_lowlmiflivaeblgbyxkxrbt3nq	109	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	109	14	conjugation	conjugation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	109	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	110	1	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	110	2	w	w	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	110	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	110	4	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	110	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	110	6	either	either	DT
work_lowlmiflivaeblgbyxkxrbt3nq	110	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	110	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	110	9	above	above	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	110	10	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	110	11	categories	category	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	110	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	110	13	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	110	14	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	110	15	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	110	16	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	110	17	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	110	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	110	19	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	110	20	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	110	21	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	110	22	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	110	23	allows	allow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	110	24	changes	change	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	110	25	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	110	26	conjugation	conjugation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	110	27	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	110	28	declension	declension	NN
work_lowlmiflivaeblgbyxkxrbt3nq	110	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	111	1	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	111	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	111	3	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	4	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	111	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	111	7	noun	noun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	8	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	111	9	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	111	10	swap	swap	VB
work_lowlmiflivaeblgbyxkxrbt3nq	111	11	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	111	12	singular	singular	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	111	13	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	111	14	plural	plural	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	111	15	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	16	con-	con-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	17	ditioned	ditione	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	111	18	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	111	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	111	20	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	21	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	111	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	112	1	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	112	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	112	3	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	112	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	112	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	112	6	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	112	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	112	8	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	112	9	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	112	10	swap	swap	VB
work_lowlmiflivaeblgbyxkxrbt3nq	112	11	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	112	12	any	any	DT
work_lowlmiflivaeblgbyxkxrbt3nq	112	13	conjugated	conjugated	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	112	14	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	112	15	non	non	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	112	16	-	-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	112	17	finite	finite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	112	18	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	112	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	112	20	again	again	RB
work_lowlmiflivaeblgbyxkxrbt3nq	112	21	conditioned	condition	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	112	22	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	112	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	112	24	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	112	25	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	112	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	113	1	Lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	113	2	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	113	3	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	113	4	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	113	5	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	113	6	marked	mark	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	113	7	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	113	8	CELEX	CELEX	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	113	9	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	113	10	Baayen	Baayen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	113	11	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	113	12	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	113	13	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	113	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	113	15	1996	1996	CD
work_lowlmiflivaeblgbyxkxrbt3nq	113	16	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	113	17	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	113	18	nouns	noun	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	113	19	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	113	20	verbs	verb	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	113	21	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	113	22	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	113	23	swap	swap	VB
work_lowlmiflivaeblgbyxkxrbt3nq	113	24	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	113	25	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	113	26	exact	exact	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	113	27	same	same	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	113	28	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	113	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	114	1	Overgeneralization	overgeneralization	NN
work_lowlmiflivaeblgbyxkxrbt3nq	114	2	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	114	3	lines	line	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	114	4	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	114	5	-	-	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	114	6	32	32	CD
work_lowlmiflivaeblgbyxkxrbt3nq	114	7	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	114	8	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	114	9	Finally	finally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	114	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	114	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	114	12	noisy	noisy	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	114	13	channel	channel	NN
work_lowlmiflivaeblgbyxkxrbt3nq	114	14	considers	consider	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	114	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	114	16	possibility	possibility	NN
work_lowlmiflivaeblgbyxkxrbt3nq	114	17	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	114	18	produc-	produc-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	114	19	ing	e	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	114	20	overgeneralized	overgeneralize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	114	21	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	114	22	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	114	23	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	114	24	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	114	25	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	114	26	maked	maked	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	114	27	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	114	28	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	114	29	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	114	30	childs	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	114	31	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	114	32	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	114	33	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	114	34	place	place	NN
work_lowlmiflivaeblgbyxkxrbt3nq	114	35	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	114	36	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	114	37	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	114	38	irregular	irregular	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	114	39	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	114	40	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	115	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	115	2	OVERGEN	OVERGEN	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	115	3	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	115	4	produces	produce	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	115	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	115	6	incorrect	incorrect	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	115	7	over-	over-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	115	8	generalized	generalize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	115	9	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	115	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	116	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	116	2	draw	draw	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	116	3	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	116	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	116	5	distribution	distribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	116	6	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	116	7	chooses	choose	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	116	8	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	116	9	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	116	10	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	116	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	116	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	116	13	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	116	14	original	original	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	116	15	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	116	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	117	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	117	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	117	3	maintains	maintain	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	117	4	separate	separate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	117	5	dis-	dis-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	117	6	tributions	tribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	117	7	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	117	8	nouns	noun	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	117	9	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	117	10	overgeneralized	overgeneralize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	117	11	plurals	plural	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	117	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	117	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	117	14	verbs	verb	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	117	15	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	117	16	overgeneralized	overgeneralize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	117	17	past	past	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	117	18	tense	tense	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	117	19	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	117	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	118	1	5	5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	118	2	Implementation	implementation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	118	3	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	118	4	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	118	5	section	section	NN
work_lowlmiflivaeblgbyxkxrbt3nq	118	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	118	7	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	118	8	describe	describe	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	118	9	steps	step	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	118	10	necessary	necessary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	118	11	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	118	12	build	build	VB
work_lowlmiflivaeblgbyxkxrbt3nq	118	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	118	14	train	train	VB
work_lowlmiflivaeblgbyxkxrbt3nq	118	15	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	118	16	test	test	VB
work_lowlmiflivaeblgbyxkxrbt3nq	118	17	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	118	18	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	118	19	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	118	20	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	118	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	119	1	Weighted	Weighted	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	2	Finite	Finite	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	3	State	State	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	4	Transducers	Transducers	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	119	5	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	119	6	FSTs	FSTs	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	7	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	119	8	used	use	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	119	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	119	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	119	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	119	12	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	119	13	constructed	construct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	119	14	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	119	15	OpenFst	OpenFst	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	16	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	119	17	Allauzen	Allauzen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	18	et	et	FW
work_lowlmiflivaeblgbyxkxrbt3nq	119	19	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	20	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	119	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	119	22	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	119	23	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	119	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	120	1	5.1	5.1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	120	2	Sentence	Sentence	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	120	3	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	120	4	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	120	5	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	120	6	provide	provide	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	120	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	120	8	basis	basis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	120	9	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	120	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	120	11	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	120	12	pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	120	13	cess	cess	NN
work_lowlmiflivaeblgbyxkxrbt3nq	120	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	121	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	121	2	represent	represent	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	121	3	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	121	4	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	121	5	building	build	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	121	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	121	7	simple	simple	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	121	8	linear	linear	NN
work_lowlmiflivaeblgbyxkxrbt3nq	121	9	chain	chain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	121	10	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	121	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	121	12	progressing	progress	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	121	13	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	121	14	node	node	RB
work_lowlmiflivaeblgbyxkxrbt3nq	121	15	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	121	16	node	node	VB
work_lowlmiflivaeblgbyxkxrbt3nq	121	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	121	18	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	121	19	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	121	20	accepting	accept	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	121	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	121	22	yielding	yield	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	121	23	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	121	24	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	121	25	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	121	26	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	121	27	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	121	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	122	1	All	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	122	2	arcs	arc	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	122	3	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	122	4	weighted	weight	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	122	5	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	122	6	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	122	7	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	122	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	123	1	4Auxiliary	4auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	123	2	lemmas	lemmas	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	123	3	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	123	4	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	123	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	123	6	do	do	VB
work_lowlmiflivaeblgbyxkxrbt3nq	123	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	123	8	go	go	VB
work_lowlmiflivaeblgbyxkxrbt3nq	123	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	123	10	will	will	MD
work_lowlmiflivaeblgbyxkxrbt3nq	123	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	123	12	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	123	13	get	get	VB
work_lowlmiflivaeblgbyxkxrbt3nq	123	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	124	1	5.2	5.2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	124	2	Noise	Noise	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	124	3	FST	fst	IN
work_lowlmiflivaeblgbyxkxrbt3nq	124	4	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	124	5	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	7	provides	provide	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	124	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	124	9	conditional	conditional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	124	10	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	11	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	124	12	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	13	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	124	14	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	124	15	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	124	16	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	17	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	124	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	125	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	125	2	en-	en-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	125	3	code	code	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	125	4	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	125	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	125	6	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	125	7	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	125	8	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	125	9	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	125	10	several	several	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	125	11	states	state	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	125	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	125	13	allow-	allow-	CC
work_lowlmiflivaeblgbyxkxrbt3nq	125	14	ing	e	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	125	15	us	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	125	16	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	125	17	track	track	VB
work_lowlmiflivaeblgbyxkxrbt3nq	125	18	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	125	19	number	number	NN
work_lowlmiflivaeblgbyxkxrbt3nq	125	20	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	125	21	consecutive	consecutive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	125	22	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	125	23	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	125	24	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	125	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	126	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	126	2	allow	allow	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	126	3	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	126	4	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	126	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	126	6	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	126	7	operations	operation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	126	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	126	9	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	126	10	row	row	NN
work_lowlmiflivaeblgbyxkxrbt3nq	126	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	126	12	thereby	thereby	RB
work_lowlmiflivaeblgbyxkxrbt3nq	126	13	constraining	constrain	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	126	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	126	15	length	length	NN
work_lowlmiflivaeblgbyxkxrbt3nq	126	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	126	17	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	126	18	out-	out-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	126	19	put	put	NN
work_lowlmiflivaeblgbyxkxrbt3nq	126	20	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	126	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	127	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	127	2	constraint	constraint	NN
work_lowlmiflivaeblgbyxkxrbt3nq	127	3	results	result	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	127	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	127	5	three	three	CD
work_lowlmiflivaeblgbyxkxrbt3nq	127	6	states	state	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	127	7	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	127	8	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	127	9	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	127	10	0	0	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	127	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	127	12	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	127	13	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	127	14	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	127	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	127	16	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	127	17	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	127	18	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	127	19	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	127	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	127	21	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	127	22	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	127	23	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	127	24	end	end	NN
work_lowlmiflivaeblgbyxkxrbt3nq	127	25	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	127	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	128	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	128	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	128	3	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	128	4	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	128	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	128	6	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	128	7	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	128	8	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	128	9	pairs	pair	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	128	10	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	128	11	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	128	12	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	128	13	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	128	14	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	128	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	128	16	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	128	17	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	128	18	due	due	IN
work_lowlmiflivaeblgbyxkxrbt3nq	128	19	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	128	20	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	128	21	constraint	constraint	NN
work_lowlmiflivaeblgbyxkxrbt3nq	128	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	129	1	Each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	129	2	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	129	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	129	5	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	129	6	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	129	7	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	129	8	�	�	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	129	9	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	129	10	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	129	12	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	13	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	14	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	129	15	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	16	symbol	symbol	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	129	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	129	19	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	129	20	possibly	possibly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	129	21	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	129	22	child-	child-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	23	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	24	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	25	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	129	26	�	�	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	129	27	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	129	28	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	29	symbol	symbol	NN
work_lowlmiflivaeblgbyxkxrbt3nq	129	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	130	1	Each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	130	2	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	3	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	130	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	130	6	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	130	8	transducing	transduce	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	130	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	130	10	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	11	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	12	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	130	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	130	14	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	15	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	130	17	determined	determine	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	130	18	according	accord	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	130	19	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	130	20	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	130	21	parameterized	parameterize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	130	22	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	130	23	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	130	24	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	130	25	Sec-	Sec-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	130	26	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	130	27	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	130	28	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	131	1	Arcs	arcs	DT
work_lowlmiflivaeblgbyxkxrbt3nq	131	2	corresponding	correspond	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	131	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	131	4	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	131	5	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	131	6	dele-	dele-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	131	7	tions	tion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	131	8	lead	lead	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	131	9	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	131	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	131	11	new	new	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	131	12	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	131	13	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	131	14	insdel++	insdel++	NN
work_lowlmiflivaeblgbyxkxrbt3nq	131	15	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	131	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	131	17	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	131	18	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	131	19	al-	al-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	131	20	lowed	low	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	131	21	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	131	22	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	131	23	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	131	24	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	131	25	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	131	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	132	1	Substitution	substitution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	132	2	arcs	arc	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	132	3	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	132	4	lead	lead	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	132	5	back	back	RB
work_lowlmiflivaeblgbyxkxrbt3nq	132	6	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	132	7	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	132	8	insdel	insdel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	132	9	=	=	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	132	10	0	0	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	132	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	133	1	Word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	2	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	3	infor-	infor-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	4	mation	mation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	133	6	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	133	7	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	133	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	133	9	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	133	11	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	12	lists	list	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	133	13	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	133	14	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	133	15	non-	non-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	16	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	133	17	class.5	class.5	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	133	18	Inflectional	inflectional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	133	19	information	information	NN
work_lowlmiflivaeblgbyxkxrbt3nq	133	20	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	133	21	derived	derive	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	133	22	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	133	23	CELEX	CELEX	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	133	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	134	1	5.3	5.3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	134	2	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	134	3	Model	Model	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	134	4	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	134	5	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	134	6	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	134	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	134	8	provides	provide	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	134	9	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	134	10	prior	prior	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	134	11	distribution	distribution	NN
work_lowlmiflivaeblgbyxkxrbt3nq	134	12	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	134	13	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	134	14	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	134	15	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	134	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	135	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	135	2	build	build	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	135	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	135	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	135	5	trigram	trigram	NN
work_lowlmiflivaeblgbyxkxrbt3nq	135	6	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	135	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	135	8	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	9	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	135	10	Kneser	Kneser	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	135	12	Ney	Ney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	13	smoothing	smooth	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	135	14	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	135	15	OpenGRM	opengrm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	135	16	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	135	17	Roark	Roark	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	18	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	19	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	20	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	135	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	135	22	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	135	23	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	135	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	136	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	136	2	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	136	3	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	136	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	136	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	136	6	trained	train	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	136	7	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	136	8	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	136	9	parent	parent	NN
work_lowlmiflivaeblgbyxkxrbt3nq	136	10	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	136	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	136	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	136	13	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	136	14	studies	study	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	136	15	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	136	16	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	136	17	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	136	18	errorful	errorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	136	19	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	136	20	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	136	21	drawn	draw	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	136	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	137	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	137	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	137	3	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	137	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	137	5	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	137	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	137	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	137	8	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	137	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	137	10	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	137	11	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	137	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	137	13	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	137	14	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	137	15	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	137	16	identical	identical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	137	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	138	1	Arcs	arc	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	138	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	138	3	weighted	weight	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	138	4	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	138	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	138	6	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	138	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	138	9	n	n	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	138	11	gram	gram	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	12	beginning	beginning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	13	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	138	14	some	some	DT
work_lowlmiflivaeblgbyxkxrbt3nq	138	15	prefix	prefix	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	16	associated	associate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	138	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	138	18	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	138	19	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	20	node	node	RB
work_lowlmiflivaeblgbyxkxrbt3nq	138	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	138	22	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	138	23	ending	end	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	138	24	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	138	25	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	138	26	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	27	’s	’s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	138	28	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	29	/	/	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	138	30	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	31	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	138	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	139	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	139	2	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	3	setup	setup	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	139	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	6	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	139	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	9	string	string	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	10	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	139	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	12	total	total	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	139	13	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	139	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	16	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	17	accepting	accept	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	139	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	139	19	emitting	emit	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	139	20	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	139	21	string	string	NN
work_lowlmiflivaeblgbyxkxrbt3nq	139	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	140	1	5.4	5.4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	140	2	Training	Training	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	140	3	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	4	detailed	detail	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	140	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	6	Section	section	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	7	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	140	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	140	9	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	140	10	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	12	consists	consist	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	140	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	140	15	series	series	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	17	multinomial	multinomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	140	18	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	140	19	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	140	20	govern	govern	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	140	21	5Word	5Word	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	140	22	lists	list	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	140	23	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	140	24	included	include	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	140	25	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	26	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	27	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	140	28	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	140	29	dataset	dataset	NN
work_lowlmiflivaeblgbyxkxrbt3nq	140	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	141	1	129	129	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	2	0	0	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	3	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	4	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	141	5	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	6	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	141	7	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	8	is:<eps	is:<ep	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	141	9	>	>	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	141	10	3him	3him	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	11	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	12	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	141	13	his	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	141	14	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	15	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	141	16	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	141	17	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	18	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	141	19	his	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	141	20	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	21	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	141	22	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	23	hat	hat	NN
work_lowlmiflivaeblgbyxkxrbt3nq	141	24	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	25	hat	hat	NN
work_lowlmiflivaeblgbyxkxrbt3nq	141	26	hats	hat	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	141	27	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	28	hat	hat	NN
work_lowlmiflivaeblgbyxkxrbt3nq	141	29	5	5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	141	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	141	31	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	141	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	142	1	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	2	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	142	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	142	4	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	142	5	simplified	simplified	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	142	6	decoding	decode	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	142	7	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	142	8	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	142	10	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	11	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	12	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	142	13	That	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	14	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	142	15	hat	hat	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	142	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	142	17	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	142	18	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	19	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	142	20	actual	actual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	142	21	decoding	decode	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	142	22	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	142	23	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	142	24	more	more	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	142	25	transduction	transduction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	26	arcs	arc	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	142	27	exist	exist	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	142	28	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	142	29	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	142	30	those	those	DT
work_lowlmiflivaeblgbyxkxrbt3nq	142	31	translating	translating	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	32	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	142	33	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	34	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	142	35	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	142	36	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	142	37	him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	142	38	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	142	39	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	40	any	any	DT
work_lowlmiflivaeblgbyxkxrbt3nq	142	41	determiner	determiner	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	42	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	142	43	pronoun	pronoun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	142	44	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	142	45	respectively	respectively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	142	46	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	142	47	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	142	48	affording	afford	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	142	49	opportunities	opportunity	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	142	50	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	142	51	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	142	52	more	more	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	142	53	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	142	54	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	142	55	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	142	56	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	143	1	Input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	143	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	143	3	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	143	4	strings	string	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	143	5	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	143	6	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	143	7	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	143	8	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	143	9	correspond	correspond	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	143	10	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	143	11	possible	possible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	143	12	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	143	13	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	143	14	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	143	15	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	143	16	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	143	17	translations	translation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	143	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	144	1	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	144	2	transformation	transformation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	144	3	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	144	4	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	144	5	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	144	6	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	144	7	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	144	8	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	144	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	144	10	al-	al-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	144	11	lowing	low	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	144	12	limited	limited	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	144	13	insertions	insertion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	144	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	144	15	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	144	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	145	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	145	2	estimate	estimate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	145	3	parameters	parameter	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	145	4	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	145	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	145	6	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	145	7	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	145	8	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	145	9	maximize	maximize	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	145	10	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	145	11	posterior	posterior	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	145	12	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	145	13	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	145	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	145	15	observed	observed	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	145	16	train-	train-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	145	17	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	145	18	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	145	19	{	{	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	145	20	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	145	21	s	s	NN
work_lowlmiflivaeblgbyxkxrbt3nq	145	22	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	145	23	t	t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	145	24	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	145	25	}	}	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	145	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	146	1	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	146	3	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	5	P(t	p(t	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	6	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	146	7	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	146	8	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	146	9	depend	depend	VB
work_lowlmiflivaeblgbyxkxrbt3nq	146	10	on	on	RP
work_lowlmiflivaeblgbyxkxrbt3nq	146	11	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	13	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	14	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	15	param-	param-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	16	eters	eter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	18	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	19	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	20	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	146	21	equivalent	equivalent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	22	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	146	23	jointly	jointly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	146	24	maximiz-	maximiz-	VB
work_lowlmiflivaeblgbyxkxrbt3nq	146	25	ing	e	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	146	26	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	27	prior	prior	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	28	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	146	29	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	30	conditional	conditional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	31	likelihoods	likelihood	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	32	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	33	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	34	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	35	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	146	36	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	37	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	38	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	146	39	argmax	argmax	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	40	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	41	P(θ	P(θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	42	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	146	43	∏	∏	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	44	P(s|t	P(s|t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	45	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	46	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	47	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	146	48	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	146	49	represent	represent	VB
work_lowlmiflivaeblgbyxkxrbt3nq	146	50	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	51	possible	possible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	52	derivations	derivation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	53	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	54	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	55	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	56	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	57	s	s	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	146	58	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	59	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	146	60	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	61	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	62	t	t	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	63	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	64	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	146	65	compose	compose	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	146	66	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	67	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	68	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	69	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	70	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	71	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	72	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	73	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	74	obtaining	obtain	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	146	75	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	146	76	FSTtrain	fsttrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	77	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	146	78	FSTt	FSTt	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	79	◦	◦	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	80	FSTnoise	fstnoise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	81	◦	◦	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	82	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	146	83	Each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	84	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	85	through	through	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	86	FSTtrain	fsttrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	87	corresponds	correspond	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	146	88	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	89	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	146	90	sin-	sin-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	91	gle	gle	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	146	92	derivation	derivation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	93	d	d	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	94	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	95	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	146	96	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	97	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	146	98	P(s	p(s	CD
work_lowlmiflivaeblgbyxkxrbt3nq	146	99	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	100	d|t	d|t	CD
work_lowlmiflivaeblgbyxkxrbt3nq	146	101	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	146	102	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	146	103	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	146	104	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	147	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	147	2	summing	sum	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	147	3	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	147	4	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	147	5	weights	weight	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	147	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	147	7	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	147	8	obtain	obtain	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	147	9	P(s|t	P(s|t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	147	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	147	11	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	147	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	147	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	148	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	148	2	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	148	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	148	4	MAP	MAP	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	148	5	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	148	6	EM	em	NN
work_lowlmiflivaeblgbyxkxrbt3nq	148	7	algorithm	algorithm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	148	8	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	148	9	maximize	maximize	VB
work_lowlmiflivaeblgbyxkxrbt3nq	148	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	148	11	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	148	12	while	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	148	13	summing	sum	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	148	14	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	148	15	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	148	16	possible	possible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	148	17	derivations	derivation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	148	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	149	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	149	2	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	3	scheme	scheme	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	4	relies	rely	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	149	5	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	149	6	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	149	7	weighted	weight	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	149	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	149	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	149	10	V	v	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	149	12	expectation	expectation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	13	semiring	semiring	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	14	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	149	15	Eisner	Eisner	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	149	17	2001	2001	CD
work_lowlmiflivaeblgbyxkxrbt3nq	149	18	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	149	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	149	20	imple-	imple-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	21	mented	mente	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	149	22	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	149	23	code	code	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	24	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	149	25	fstrain	fstrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	149	26	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	149	27	Dreyer	Dreyer	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	28	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	29	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	30	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	149	31	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	149	32	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	149	33	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	149	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	150	1	Besides	besides	IN
work_lowlmiflivaeblgbyxkxrbt3nq	150	2	carrying	carry	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	150	3	probabilities	probability	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	150	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	150	5	arc	arc	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	150	6	weights	weight	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	150	7	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	150	8	sup-	sup-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	150	9	plemented	plemente	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	150	10	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	150	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	150	12	vector	vector	NN
work_lowlmiflivaeblgbyxkxrbt3nq	150	13	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	150	14	indicate	indicate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	150	15	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	150	16	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	150	17	involved	involve	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	150	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	150	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	150	20	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	150	21	traversal	traversal	NN
work_lowlmiflivaeblgbyxkxrbt3nq	150	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	151	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	151	2	V	V	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	151	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	151	4	expectation	expectation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	151	5	semiring	semiring	NN
work_lowlmiflivaeblgbyxkxrbt3nq	151	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	151	7	designed	design	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	151	8	so	so	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	9	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	151	11	total	total	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	151	12	arc	arc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	151	13	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	151	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	15	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	151	16	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	151	17	through	through	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	18	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	151	19	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	151	20	yields	yield	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	151	21	both	both	CC
work_lowlmiflivaeblgbyxkxrbt3nq	151	22	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	151	23	probabil-	probabil-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	151	24	ity	ity	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	151	25	P(s|t	P(s|t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	151	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	151	27	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	151	28	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	151	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	151	30	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	31	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	151	32	expected	expect	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	151	33	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	151	34	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	151	35	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	152	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	152	2	EM	EM	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	152	3	algorithm	algorithm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	152	4	proceeds	proceed	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	152	5	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	152	6	follows	follow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	152	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	152	8	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	152	9	start	start	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	152	10	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	152	11	initializing	initialize	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	152	12	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	152	13	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	152	14	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	152	15	uniform	uniform	VB
work_lowlmiflivaeblgbyxkxrbt3nq	152	16	distribu-	distribu-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	152	17	tions	tion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	152	18	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	152	19	random	random	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	152	20	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	152	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	153	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	153	2	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	153	3	weight	weight	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	153	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	153	5	arcs	arc	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	153	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	153	7	FSTnoise	FSTnoise	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	153	8	accordingly	accordingly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	153	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	154	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	154	2	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	154	3	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	4	pair	pair	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	5	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	154	6	s	s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	154	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	154	8	t	t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	154	9	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	154	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	154	11	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	154	12	build	build	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	154	13	FSTtrain	fsttrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	14	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	154	15	composition	composition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	16	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	154	17	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	154	18	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	19	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	154	21	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	154	22	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	154	23	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	154	24	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	154	25	previous	previous	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	154	26	paragraph	paragraph	NN
work_lowlmiflivaeblgbyxkxrbt3nq	154	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	155	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	155	2	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	155	3	compute	compute	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	155	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	155	5	total	total	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	155	6	arc	arc	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	155	7	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	155	9	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	155	10	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	155	11	through	through	IN
work_lowlmiflivaeblgbyxkxrbt3nq	155	12	FSTtrain	fsttrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	13	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	155	14	relabeling	relabele	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	155	15	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	155	16	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	17	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	155	18	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	19	symbols	symbol	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	155	20	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	155	21	�	�	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	155	22	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	155	23	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	155	24	reducing	reduce	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	155	25	FSTtrain	fsttrain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	26	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	155	27	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	155	28	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	155	29	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	30	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	155	31	epsilon	epsilon	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	32	removal	removal	NN
work_lowlmiflivaeblgbyxkxrbt3nq	155	33	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	155	34	Mohri	Mohri	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	155	35	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	155	36	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	155	37	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	155	38	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	156	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	2	stop-	stop-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	156	3	ping	ping	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	156	4	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	156	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	6	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	7	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	156	8	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	156	9	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	156	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	11	sum	sum	NN
work_lowlmiflivaeblgbyxkxrbt3nq	156	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	13	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	14	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	156	15	through	through	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	17	original	original	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	156	18	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	156	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	156	20	yielding	yield	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	156	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	156	22	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	156	23	P(s|t	P(s|t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	156	24	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	156	25	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	156	26	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	156	27	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	156	28	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	29	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	30	expected	expect	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	156	31	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	156	32	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	156	33	ac-	ac-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	156	34	cording	cord	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	156	35	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	156	36	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	156	37	current	current	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	156	38	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	156	39	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	157	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	157	2	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	157	3	reesti-	reesti-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	157	4	mate	mate	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	157	5	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	157	6	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	157	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	157	8	expected	expect	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	157	9	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	157	10	plus	plus	CC
work_lowlmiflivaeblgbyxkxrbt3nq	157	11	pseudo	pseudo	NN
work_lowlmiflivaeblgbyxkxrbt3nq	157	12	-	-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	157	13	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	157	14	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	157	15	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	157	16	priors	prior	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	157	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	157	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	157	19	repeat	repeat	VB
work_lowlmiflivaeblgbyxkxrbt3nq	157	20	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	157	21	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	157	22	until	until	IN
work_lowlmiflivaeblgbyxkxrbt3nq	157	23	conver-	conver-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	157	24	gence	gence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	157	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	158	1	Besides	besides	IN
work_lowlmiflivaeblgbyxkxrbt3nq	158	2	smoothing	smooth	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	158	3	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	158	4	estimated	estimate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	158	5	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	158	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	158	8	pseudo	pseudo	NN
work_lowlmiflivaeblgbyxkxrbt3nq	158	9	-	-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	158	10	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	11	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	158	12	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	158	13	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	158	14	asymmetric	asymmetric	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	158	15	Dirich-	Dirich-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	158	16	let	let	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	158	17	priors	prior	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	18	favor	favor	VB
work_lowlmiflivaeblgbyxkxrbt3nq	158	19	multinomials	multinomial	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	20	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	158	21	retain	retain	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	158	22	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	158	23	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	158	24	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	158	25	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	158	26	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	158	27	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	28	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	158	29	identical	identical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	158	30	lemmas	lemma	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	31	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	158	32	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	158	33	identical	identical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	158	34	inflections	inflection	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	158	35	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	158	36	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	159	1	Concretely	concretely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	159	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	159	3	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	159	4	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	159	5	pseudo	pseudo	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	6	-	-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	7	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	9	.5	.5	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	10	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	11	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	159	12	favored	favor	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	159	13	outcomes	outcome	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	159	15	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	159	16	pseudo	pseudo	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	17	-	-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	18	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	20	.01	.01	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	159	21	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	22	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	159	23	others.6	others.6	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	24	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	25	practice	practice	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	159	27	109	109	CD
work_lowlmiflivaeblgbyxkxrbt3nq	159	28	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	29	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	159	30	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	31	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	159	32	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	33	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	159	34	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	35	set	set	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	159	36	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	159	37	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	159	38	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	159	39	translated	translate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	159	40	into	into	IN
work_lowlmiflivaeblgbyxkxrbt3nq	159	41	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	159	42	corresponding	corresponding	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	159	43	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	44	version	version	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	45	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	159	46	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	159	47	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	159	48	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	160	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	160	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	160	3	due	due	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	160	4	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	160	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	160	6	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	160	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	160	8	rare	rare	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	160	9	phenomena	phenomenon	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	160	10	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	160	11	rephrasing	rephrasing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	160	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	160	13	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	160	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	160	15	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	160	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	160	17	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	160	18	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	160	19	order	order	NN
work_lowlmiflivaeblgbyxkxrbt3nq	160	20	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	160	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	161	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	161	2	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	161	3	cases	case	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	161	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	161	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	161	6	composed	compose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	161	7	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	161	8	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	161	9	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	161	10	valid	valid	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	161	11	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	161	12	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	161	13	start	start	NN
work_lowlmiflivaeblgbyxkxrbt3nq	161	14	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	161	15	finish	finish	NN
work_lowlmiflivaeblgbyxkxrbt3nq	161	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	161	17	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	161	18	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	161	19	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	161	20	removed	remove	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	161	21	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	161	22	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	161	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	162	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	162	2	run	run	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	162	3	EM	em	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	162	4	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	162	5	100	100	CD
work_lowlmiflivaeblgbyxkxrbt3nq	162	6	iterations	iteration	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	162	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	162	8	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	162	9	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	162	10	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	162	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	162	12	log	log	NN
work_lowlmiflivaeblgbyxkxrbt3nq	162	13	likelihood	likelihood	NN
work_lowlmiflivaeblgbyxkxrbt3nq	162	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	162	15	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	162	16	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	162	17	generally	generally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	162	18	converges	converge	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	162	19	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	162	20	within	within	RB
work_lowlmiflivaeblgbyxkxrbt3nq	162	21	.01	.01	CD
work_lowlmiflivaeblgbyxkxrbt3nq	162	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	163	1	5.5	5.5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	163	2	Decoding	decode	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	163	3	After	after	IN
work_lowlmiflivaeblgbyxkxrbt3nq	163	4	training	train	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	163	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	163	6	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	163	9	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	163	10	apply	apply	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	163	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	163	12	sys-	sys-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	13	tem	tem	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	14	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	163	15	translate	translate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	163	16	divergent	divergent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	163	17	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	18	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	19	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	163	20	adult-	adult-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	163	21	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	163	22	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	163	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	164	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	2	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	3	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	164	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	164	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	164	6	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	164	7	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	164	8	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	164	9	com-	com-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	164	10	posed	pose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	164	11	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	164	13	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	164	14	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	15	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	164	16	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	164	17	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	164	18	s.	s.	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	164	19	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	20	6corresponding	6corresponding	CD
work_lowlmiflivaeblgbyxkxrbt3nq	164	21	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	22	Dirichlet	Dirichlet	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	164	23	hyperparameters	hyperparameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	164	24	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	164	25	1.5	1.5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	164	26	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	164	27	1.01	1.01	CD
work_lowlmiflivaeblgbyxkxrbt3nq	164	28	respectively	respectively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	164	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	165	1	130	130	CD
work_lowlmiflivaeblgbyxkxrbt3nq	165	2	place	place	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	165	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	165	5	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	6	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	165	9	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	10	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	11	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	12	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	165	13	used	use	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	165	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	15	yielding	yield	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	165	16	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	165	17	FSTdecode	FSTdecode	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	18	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	165	19	FSTlm	FSTlm	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	20	◦	◦	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	21	FSTnoise	FSTnoise	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	22	◦	◦	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	23	FSTs	fst	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	165	24	Each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	165	25	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	26	through	through	IN
work_lowlmiflivaeblgbyxkxrbt3nq	165	27	FSTdecode	FSTdecode	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	28	corresponds	correspond	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	165	29	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	165	30	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	165	31	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	32	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	33	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	165	34	derivation	derivation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	35	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	165	36	t	t	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	37	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	38	d	d	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	39	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	165	40	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	41	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	165	42	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	43	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	165	44	P(s	p(s	CD
work_lowlmiflivaeblgbyxkxrbt3nq	165	45	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	46	d|t	d|t	CD
work_lowlmiflivaeblgbyxkxrbt3nq	165	47	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	165	48	θ)P(t	θ)P(t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	165	49	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	165	50	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	166	1	Thus	thus	RB
work_lowlmiflivaeblgbyxkxrbt3nq	166	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	4	highest	high	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	166	5	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	166	6	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	7	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	8	corresponds	correspond	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	166	9	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	166	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	11	most	most	RBS
work_lowlmiflivaeblgbyxkxrbt3nq	166	12	likely	likely	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	166	13	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	166	15	derivation	derivation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	16	pair	pair	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	17	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	166	18	argmax	argmax	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	19	t	t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	21	d	d	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	22	P(t	P(t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	24	d|s	d|s	CD
work_lowlmiflivaeblgbyxkxrbt3nq	166	25	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	26	θ	θ	LS
work_lowlmiflivaeblgbyxkxrbt3nq	166	27	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	166	28	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	166	29	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	166	30	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	31	dynamic	dynamic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	166	32	program	program	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	33	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	166	34	find	find	VB
work_lowlmiflivaeblgbyxkxrbt3nq	166	35	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	36	n	n	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	37	highest-	highest-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	166	38	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	39	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	166	40	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	166	41	distinct	distinct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	166	42	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	43	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	166	44	t.	t.	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	45	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	46	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	166	47	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	166	48	viewed	view	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	166	49	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	166	50	finding	find	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	166	51	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	52	n	n	RB
work_lowlmiflivaeblgbyxkxrbt3nq	166	53	most	most	RBS
work_lowlmiflivaeblgbyxkxrbt3nq	166	54	likely	likely	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	166	55	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	56	trans-	trans-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	57	lations	lation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	166	58	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	59	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	166	60	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	166	61	Viterbi	Viterbi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	62	approximation	approximation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	166	63	P(t|s	P(t|s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	64	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	65	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	66	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	166	67	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	166	68	argmaxd	argmaxd	FW
work_lowlmiflivaeblgbyxkxrbt3nq	166	69	P(t	P(t	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	70	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	71	d|s	d|s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	72	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	166	73	θ	θ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	166	74	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	166	75	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	167	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	167	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	167	3	experiments	experiment	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	167	4	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	167	5	set	set	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	167	6	n	n	CC
work_lowlmiflivaeblgbyxkxrbt3nq	167	7	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	167	8	50	50	CD
work_lowlmiflivaeblgbyxkxrbt3nq	167	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	168	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	168	2	simplified	simplified	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	168	3	FSTdecode	FSTdecode	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	168	4	example	example	NN
work_lowlmiflivaeblgbyxkxrbt3nq	168	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	168	6	shown	show	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	168	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	168	8	Figure	Figure	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	168	9	1	1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	168	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	169	1	5.6	5.6	CD
work_lowlmiflivaeblgbyxkxrbt3nq	169	2	Discriminative	Discriminative	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	169	3	Reranking	Reranking	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	169	4	To	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	169	5	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	169	6	flexibly	flexibly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	169	7	capture	capture	VB
work_lowlmiflivaeblgbyxkxrbt3nq	169	8	long	long	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	169	9	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	169	10	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	169	11	fea-	fea-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	169	12	tures	ture	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	169	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	169	14	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	169	15	embed	embe	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	169	16	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	169	17	noisy	noisy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	169	18	channel	channel	NN
work_lowlmiflivaeblgbyxkxrbt3nq	169	19	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	169	20	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	169	21	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	169	22	dis-	dis-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	169	23	criminative	criminative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	169	24	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	169	25	procedure	procedure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	169	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	170	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	170	2	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	170	3	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	170	4	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	170	5	tence	tence	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	170	6	s	s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	170	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	170	8	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	170	9	take	take	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	170	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	170	11	n	n	RB
work_lowlmiflivaeblgbyxkxrbt3nq	170	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	170	13	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	170	14	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	170	15	translations	translation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	170	16	t1	t1	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	170	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	170	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	171	1	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	172	1	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	173	1	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	173	2	tn	tn	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	173	3	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	173	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	173	5	underlying	underlie	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	173	6	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	173	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	173	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	173	9	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	173	10	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	173	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	173	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	173	13	previous	previous	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	173	14	section	section	NN
work_lowlmiflivaeblgbyxkxrbt3nq	173	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	174	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	174	2	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	174	3	map	map	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	174	4	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	174	5	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	174	6	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	174	7	ti	ti	IN
work_lowlmiflivaeblgbyxkxrbt3nq	174	8	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	174	9	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	174	10	d	d	NN
work_lowlmiflivaeblgbyxkxrbt3nq	174	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	174	12	dimensional	dimensional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	174	13	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	174	14	vector	vector	NN
work_lowlmiflivaeblgbyxkxrbt3nq	174	15	f(s	f(s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	174	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	174	17	ti	ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	174	18	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	174	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	175	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	175	2	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	3	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	4	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	175	5	uses	use	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	175	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	175	7	d-	d-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	175	8	dimensional	dimensional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	175	9	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	10	vector	vector	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	11	λ	λ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	12	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	175	13	predict	predict	VB
work_lowlmiflivaeblgbyxkxrbt3nq	175	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	175	15	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	16	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	175	18	highest	high	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	175	19	linear	linear	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	175	20	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	21	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	175	22	t∗	t∗	ADD
work_lowlmiflivaeblgbyxkxrbt3nq	175	23	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	175	24	argmax	argmax	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	25	ti	ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	26	λ	λ	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	27	·	·	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	175	28	f(s	f(s	CD
work_lowlmiflivaeblgbyxkxrbt3nq	175	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	175	30	ti	ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	31	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	175	32	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	175	33	simulate	simulate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	175	34	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	35	conditions	condition	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	175	36	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	175	37	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	175	38	train	train	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	175	39	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	175	40	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	41	vec-	vec-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	42	tor	tor	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	43	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	175	44	n	n	CC
work_lowlmiflivaeblgbyxkxrbt3nq	175	45	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	175	46	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	175	47	lists	list	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	175	48	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	175	49	8-fold	8-fold	CD
work_lowlmiflivaeblgbyxkxrbt3nq	175	50	cross	cross	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	51	-	-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	52	validation	validation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	53	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	175	54	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	55	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	175	56	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	175	57	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	175	58	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	175	59	averaged	average	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	175	60	perceptron	perceptron	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	61	rerank-	rerank-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	62	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	63	algorithm	algorithm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	175	64	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	175	65	Collins	Collins	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	66	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	175	67	Roark	Roark	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	175	68	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	175	69	2004	2004	CD
work_lowlmiflivaeblgbyxkxrbt3nq	175	70	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	175	71	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	176	1	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	176	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	176	3	n	n	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	176	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	176	5	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	176	6	list	list	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	7	might	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	176	8	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	176	9	include	include	VB
work_lowlmiflivaeblgbyxkxrbt3nq	176	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	176	11	exact	exact	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	176	12	gold	gold	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	176	13	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	176	14	standard	standard	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	176	15	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	176	17	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	176	18	target	target	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	19	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	20	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	176	21	maximizes	maximize	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	176	22	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	176	23	evaluation	evaluation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	24	metric	metric	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	25	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	176	26	chosen	choose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	176	27	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	176	28	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	176	29	list	list	NN
work_lowlmiflivaeblgbyxkxrbt3nq	176	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	177	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	177	2	n	n	IN
work_lowlmiflivaeblgbyxkxrbt3nq	177	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	177	4	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	177	5	list	list	NN
work_lowlmiflivaeblgbyxkxrbt3nq	177	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	177	7	non	non	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	177	8	-	-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	177	9	linearly	linearly	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	177	10	separable	separable	NN
work_lowlmiflivaeblgbyxkxrbt3nq	177	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	177	12	so	so	RB
work_lowlmiflivaeblgbyxkxrbt3nq	177	13	perceptron	perceptron	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	177	14	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	177	15	iterates	iterate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	177	16	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	177	17	1000	1000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	177	18	rounds	round	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	177	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	177	20	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	177	21	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	177	22	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	177	23	terminated	terminate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	177	24	with-	with-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	177	25	out	out	IN
work_lowlmiflivaeblgbyxkxrbt3nq	177	26	converging	converge	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	177	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	178	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	178	2	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	3	function	function	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	178	4	f(s	f(s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	178	6	ti	ti	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	7	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	8	yields	yield	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	178	9	nine	nine	CD
work_lowlmiflivaeblgbyxkxrbt3nq	178	10	boolean	boolean	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	178	12	real	real	RB
work_lowlmiflivaeblgbyxkxrbt3nq	178	13	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	178	14	valued	value	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	178	15	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	178	16	derived	derive	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	178	17	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	178	18	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	19	i	i	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	20	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	178	22	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	23	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	178	24	generates	generate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	178	25	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	26	sentence	sentence	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	27	s	s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	178	28	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	178	29	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	30	adult-	adult-	.
work_lowlmiflivaeblgbyxkxrbt3nq	178	31	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	32	ti	ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	178	34	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	178	35	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	36	ii	ii	LS
work_lowlmiflivaeblgbyxkxrbt3nq	178	37	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	38	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	178	39	POS	POS	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	40	sequence	sequence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	41	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	178	42	dependency	dependency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	43	parse	parse	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	44	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	178	45	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	178	46	ti	ti	IN
work_lowlmiflivaeblgbyxkxrbt3nq	178	47	obtained	obtain	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	178	48	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	178	49	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	178	50	Stanford	Stanford	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	51	Parser	Parser	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	52	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	53	de	de	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	54	Marneffe	Marneffe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	55	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	56	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	57	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	178	58	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	178	59	2006	2006	CD
work_lowlmiflivaeblgbyxkxrbt3nq	178	60	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	178	61	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	179	1	Features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	179	2	were	be	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	179	3	se-	se-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	179	4	lected	lecte	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	179	5	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	179	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	179	7	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	179	8	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	179	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	179	10	reranking	reranke	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	179	11	held-	held-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	179	12	out	out	RP
work_lowlmiflivaeblgbyxkxrbt3nq	179	13	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	179	14	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	179	15	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	179	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	179	17	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	179	18	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	179	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	180	1	Rerank-	Rerank-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	2	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	3	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	180	4	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	180	5	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	180	6	below	below	RP
work_lowlmiflivaeblgbyxkxrbt3nq	180	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	180	8	Generative	Generative	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	9	Model	Model	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	10	Probabilities	Probabilities	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	180	11	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	180	12	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	180	13	first	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	180	14	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	180	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	180	16	joint	joint	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	180	17	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	18	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	180	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	180	20	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	21	sentence	sentence	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	22	s	s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	180	23	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	180	24	can-	can-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	25	didate	didate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	180	26	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	27	ti	ti	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	28	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	180	29	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	180	30	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	180	31	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	180	32	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	180	33	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	180	34	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	180	35	Plm(ti)Pnoise(s|ti	Plm(ti)Pnoise(s|ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	180	36	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	180	37	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	181	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	181	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	181	3	isolate	isolate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	181	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	181	5	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	6	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	7	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	181	8	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	181	11	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	12	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	181	13	prob-	prob-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	181	14	abilities	ability	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	181	15	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	181	16	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	181	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	182	1	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	182	2	both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	182	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	182	4	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	182	5	proba-	proba-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	182	6	bilities	bilitie	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	182	7	naturally	naturally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	182	8	favor	favor	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	182	9	shorter	short	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	182	10	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	182	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	182	12	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	182	13	scale	scale	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	182	14	them	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	182	15	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	182	16	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	182	17	length	length	NN
work_lowlmiflivaeblgbyxkxrbt3nq	182	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	182	19	yielding	yield	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	182	20	n	n	CC
work_lowlmiflivaeblgbyxkxrbt3nq	182	21	√	√	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	182	22	Plm(ti	Plm(ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	182	23	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	182	24	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	182	25	n	n	CC
work_lowlmiflivaeblgbyxkxrbt3nq	182	26	√	√	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	182	27	Pnoise(s|ti	Pnoise(s|ti	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	182	28	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	182	29	respectively	respectively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	182	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	183	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	183	2	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	183	3	scaling	scale	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	183	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	183	5	joint	joint	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	183	6	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	183	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	183	8	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	183	9	allow	allow	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	183	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	183	11	reranker	reranker	NN
work_lowlmiflivaeblgbyxkxrbt3nq	183	12	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	183	13	learn	learn	VB
work_lowlmiflivaeblgbyxkxrbt3nq	183	14	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	183	15	own	own	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	183	16	bias	bias	NN
work_lowlmiflivaeblgbyxkxrbt3nq	183	17	towards	towards	RB
work_lowlmiflivaeblgbyxkxrbt3nq	183	18	longer	longer	RB
work_lowlmiflivaeblgbyxkxrbt3nq	183	19	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	183	20	shorter	shorter	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	183	21	corrected	correct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	183	22	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	183	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	184	1	Contains	contain	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	184	2	Noun	Noun	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	184	3	Subject	Subject	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	184	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	184	5	Accusative	Accusative	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	184	6	Noun	Noun	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	184	7	Sub-	Sub-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	184	8	ject	ject	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	9	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	184	10	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	184	11	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	184	12	boolean	boolean	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	184	13	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	14	indicates	indicate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	184	15	whether	whether	IN
work_lowlmiflivaeblgbyxkxrbt3nq	184	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	184	17	dependency	dependency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	18	parse	parse	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	184	20	candidate	candidate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	21	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	22	ti	ti	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	23	con-	con-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	24	tains	tain	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	184	25	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	184	26	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	184	27	nsubj	nsubj	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	28	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	184	29	relation	relation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	184	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	185	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	185	2	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	185	3	indicates	indicate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	185	4	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	185	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	185	6	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	185	7	nsubj	nsubj	NN
work_lowlmiflivaeblgbyxkxrbt3nq	185	8	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	185	9	relation	relation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	185	10	exists	exist	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	185	11	where	where	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	185	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	185	13	dependent	dependent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	185	14	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	185	15	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	185	16	ac-	ac-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	185	17	cusative	cusative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	185	18	pronoun	pronoun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	185	19	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	185	20	e.g.	e.g.	RB
work_lowlmiflivaeblgbyxkxrbt3nq	186	1	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	186	2	Him	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	186	3	ate	eat	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	186	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	186	5	cookie	cookie	NN
work_lowlmiflivaeblgbyxkxrbt3nq	186	6	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	186	7	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	186	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	187	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	187	2	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	187	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	187	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	187	5	one	one	NN
work_lowlmiflivaeblgbyxkxrbt3nq	187	6	following	follow	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	187	7	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	187	8	previously	previously	RB
work_lowlmiflivaeblgbyxkxrbt3nq	187	9	been	be	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	187	10	used	use	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	187	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	187	12	classifier	classifier	NN
work_lowlmiflivaeblgbyxkxrbt3nq	187	13	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	187	14	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	187	15	detection	detection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	187	16	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	187	17	Morley	Morley	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	187	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	187	19	Prud’hommeaux	Prud’hommeaux	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	187	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	187	21	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	187	22	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	187	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	188	1	Contains	contain	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	188	2	Finite	Finite	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	188	3	Verb	Verb	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	188	4	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	188	5	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	188	6	boolean	boolean	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	188	7	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	188	8	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	188	9	true	true	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	188	10	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	188	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	188	12	POS	POS	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	188	13	tags	tag	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	188	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	188	15	ti	ti	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	188	16	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	188	17	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	188	18	finite	finite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	188	19	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	188	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	189	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	189	2	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	189	3	differentiates	differentiate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	189	4	structures	structure	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	189	5	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	189	6	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	189	7	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	189	8	am	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	189	9	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	189	10	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	189	11	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	189	12	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	189	13	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	189	14	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	189	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	189	16	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	189	17	Question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	189	18	Template	Template	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	189	19	Features	feature	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	189	20	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	189	21	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	189	22	define	define	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	189	23	tem-	tem-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	189	24	plates	plate	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	189	25	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	189	26	wh-	wh-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	189	27	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	189	28	yes	yes	UH
work_lowlmiflivaeblgbyxkxrbt3nq	189	29	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	189	30	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	189	31	questions	question	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	189	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	190	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	190	2	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	190	3	fits	fit	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	190	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	190	5	wh-	wh-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	190	6	question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	190	7	template	template	NN
work_lowlmiflivaeblgbyxkxrbt3nq	190	8	if	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	190	9	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	190	10	begins	begin	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	190	11	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	190	12	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	190	13	wh-	wh-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	190	14	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	190	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	190	16	followed	follow	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	190	17	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	190	18	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	190	19	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	190	20	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	190	21	copula	copula	VB
work_lowlmiflivaeblgbyxkxrbt3nq	190	22	verb	verb	FW
work_lowlmiflivaeblgbyxkxrbt3nq	190	23	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	190	24	e.g.	e.g.	RB
work_lowlmiflivaeblgbyxkxrbt3nq	191	1	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	191	2	Who	who	WP
work_lowlmiflivaeblgbyxkxrbt3nq	191	3	did	do	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	191	4	...	...	:
work_lowlmiflivaeblgbyxkxrbt3nq	191	5	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	191	6	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	191	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	192	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	2	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	3	fits	fit	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	192	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	5	yes	yes	UH
work_lowlmiflivaeblgbyxkxrbt3nq	192	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	192	7	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	8	template	template	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	9	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	192	10	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	192	11	begins	begin	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	192	12	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	192	13	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	14	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	192	15	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	192	16	copula	copula	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	192	17	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	192	19	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	192	20	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	21	noun	noun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	22	subject	subject	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	23	followed	follow	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	192	24	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	192	25	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	192	26	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	27	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	192	28	adjective	adjective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	192	29	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	192	30	e.g.	e.g.	RB
work_lowlmiflivaeblgbyxkxrbt3nq	193	1	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	193	2	Are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	193	3	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	193	4	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	193	5	...	...	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	193	6	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	193	7	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	193	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	194	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	194	2	include	include	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	194	3	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	194	4	boolean	boolean	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	194	5	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	194	7	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	194	9	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	10	templates	template	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	194	11	indicating	indicate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	194	12	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	194	13	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	14	tem-	tem-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	15	plate	plate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	16	match	match	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	17	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	194	18	inappropriate	inappropriate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	194	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	194	20	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	194	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	22	original	original	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	194	23	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	24	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	25	terminates	terminate	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	194	26	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	194	27	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	28	period	period	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	29	instead	instead	RB
work_lowlmiflivaeblgbyxkxrbt3nq	194	30	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	194	31	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	194	32	question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	33	mark	mark	NN
work_lowlmiflivaeblgbyxkxrbt3nq	194	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	195	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	195	2	addition	addition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	195	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	195	5	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	195	6	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	195	7	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	195	8	inappropri-	inappropri-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	195	9	ate	ate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	195	10	template	template	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	11	matches	match	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	195	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	195	13	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	195	14	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	195	15	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	195	16	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	195	17	feature	feature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	18	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	195	19	signals	signal	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	195	20	appropriate	appropriate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	195	21	matches	match	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	195	22	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	195	23	either	either	DT
work_lowlmiflivaeblgbyxkxrbt3nq	195	24	question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	25	tem-	tem-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	26	plate	plate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	27	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	195	28	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	195	29	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	195	30	original	original	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	195	31	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	32	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	33	terminates	terminate	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	195	34	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	195	35	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	195	36	question	question	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	37	mark	mark	NN
work_lowlmiflivaeblgbyxkxrbt3nq	195	38	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	196	1	131	131	CD
work_lowlmiflivaeblgbyxkxrbt3nq	196	2	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	3	Utterance	Utterance	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	4	Human	Human	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	5	Correction	Correction	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	6	Machine	Machine	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	7	Correction	Correction	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	196	8	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	196	9	am	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	196	10	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	196	11	put	put	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	196	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	196	13	my	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	196	14	mouth	mouth	NN
work_lowlmiflivaeblgbyxkxrbt3nq	196	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	197	1	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	197	2	am	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	197	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	197	4	putting	put	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	197	5	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	197	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	197	7	my	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	197	8	mouth	mouth	NN
work_lowlmiflivaeblgbyxkxrbt3nq	197	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	198	1	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	198	2	am	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	198	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	198	4	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	198	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	198	6	put	put	VB
work_lowlmiflivaeblgbyxkxrbt3nq	198	7	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	198	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	198	9	my	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	198	10	mouth	mouth	NN
work_lowlmiflivaeblgbyxkxrbt3nq	198	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	199	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	199	2	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	199	3	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	199	4	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	199	5	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	200	1	Does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	200	2	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	200	3	one	one	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	200	4	have	have	VB
work_lowlmiflivaeblgbyxkxrbt3nq	200	5	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	200	6	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	201	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	201	2	one	one	NN
work_lowlmiflivaeblgbyxkxrbt3nq	201	3	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	201	4	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	201	5	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	202	1	Want	want	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	202	2	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	202	3	read	read	VB
work_lowlmiflivaeblgbyxkxrbt3nq	202	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	202	5	book	book	NN
work_lowlmiflivaeblgbyxkxrbt3nq	202	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	203	1	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	203	2	want	want	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	203	3	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	203	4	read	read	VB
work_lowlmiflivaeblgbyxkxrbt3nq	203	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	203	6	book	book	NN
work_lowlmiflivaeblgbyxkxrbt3nq	203	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	204	1	You	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	204	2	want	want	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	204	3	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	204	4	read	read	VB
work_lowlmiflivaeblgbyxkxrbt3nq	204	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	204	6	book	book	NN
work_lowlmiflivaeblgbyxkxrbt3nq	204	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	205	1	Why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	205	2	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	205	3	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	205	4	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	205	5	get	get	VB
work_lowlmiflivaeblgbyxkxrbt3nq	205	6	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	205	7	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	206	1	Why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	206	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	206	3	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	206	4	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	206	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	206	6	get	get	VB
work_lowlmiflivaeblgbyxkxrbt3nq	206	7	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	206	8	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	207	1	Why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	207	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	207	3	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	207	4	going	go	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	207	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	207	6	have	have	VB
work_lowlmiflivaeblgbyxkxrbt3nq	207	7	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	207	8	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	208	1	You	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	208	2	very	very	RB
work_lowlmiflivaeblgbyxkxrbt3nq	208	3	sticky	sticky	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	208	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	209	1	You	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	209	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	209	3	very	very	RB
work_lowlmiflivaeblgbyxkxrbt3nq	209	4	sticky	sticky	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	209	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	210	1	You	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	210	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	210	3	very	very	RB
work_lowlmiflivaeblgbyxkxrbt3nq	210	4	sticky	sticky	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	210	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	211	1	He	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	211	2	no	no	RB
work_lowlmiflivaeblgbyxkxrbt3nq	211	3	like	like	UH
work_lowlmiflivaeblgbyxkxrbt3nq	211	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	212	1	He	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	212	2	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	212	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	212	4	like	like	VB
work_lowlmiflivaeblgbyxkxrbt3nq	212	5	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	212	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	213	1	He	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	213	2	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	213	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	213	4	like	like	VB
work_lowlmiflivaeblgbyxkxrbt3nq	213	5	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	213	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	214	1	Yeah	yeah	UH
work_lowlmiflivaeblgbyxkxrbt3nq	214	2	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	214	3	looks	look	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	214	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	214	5	lady	lady	NN
work_lowlmiflivaeblgbyxkxrbt3nq	214	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	215	1	Yeah	yeah	UH
work_lowlmiflivaeblgbyxkxrbt3nq	215	2	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	215	3	looks	look	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	215	4	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	215	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	215	6	lady	lady	NN
work_lowlmiflivaeblgbyxkxrbt3nq	215	7	Yeah	yeah	UH
work_lowlmiflivaeblgbyxkxrbt3nq	215	8	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	215	9	looks	look	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	215	10	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	215	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	215	12	lady	lady	NN
work_lowlmiflivaeblgbyxkxrbt3nq	215	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	216	1	Eleanor	Eleanor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	216	2	come	come	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	216	3	too	too	RB
work_lowlmiflivaeblgbyxkxrbt3nq	216	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	217	1	Eleanor	Eleanor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	217	2	came	come	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	217	3	too	too	RB
work_lowlmiflivaeblgbyxkxrbt3nq	217	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	218	1	Eleanor	Eleanor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	218	2	come	come	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	218	3	too	too	RB
work_lowlmiflivaeblgbyxkxrbt3nq	218	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	219	1	Desk	desk	NN
work_lowlmiflivaeblgbyxkxrbt3nq	219	2	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	219	3	here	here	RB
work_lowlmiflivaeblgbyxkxrbt3nq	219	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	220	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	220	2	desk	desk	NN
work_lowlmiflivaeblgbyxkxrbt3nq	220	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	220	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	220	5	here	here	RB
work_lowlmiflivaeblgbyxkxrbt3nq	220	6	Desk	Desk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	220	7	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	220	8	in	in	RB
work_lowlmiflivaeblgbyxkxrbt3nq	220	9	here	here	RB
work_lowlmiflivaeblgbyxkxrbt3nq	220	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	221	1	Why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	221	2	he	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	221	3	’s	’	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	221	4	doc	doc	VB
work_lowlmiflivaeblgbyxkxrbt3nq	221	5	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	222	1	Why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	222	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	222	3	he	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	222	4	called	call	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	222	5	doc	doc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	222	6	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	223	1	He	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	223	2	’s	’	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	223	3	up	up	RB
work_lowlmiflivaeblgbyxkxrbt3nq	223	4	doc	doc	NN
work_lowlmiflivaeblgbyxkxrbt3nq	223	5	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	224	1	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	2	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	224	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	224	4	Randomly	randomly	VB
work_lowlmiflivaeblgbyxkxrbt3nq	224	5	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	224	6	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	7	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	8	generated	generate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	224	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	224	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	224	11	complete	complete	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	224	12	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	13	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	14	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	224	16	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	224	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	224	18	corresponding	corresponding	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	224	19	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	224	20	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	224	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	224	22	human	human	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	224	23	corrections	correction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	224	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	225	1	6	6	CD
work_lowlmiflivaeblgbyxkxrbt3nq	225	2	Experiments	Experiments	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	225	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	225	4	Analysis	Analysis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	225	5	Baselines	Baselines	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	225	6	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	225	7	compare	compare	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	225	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	225	9	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	225	10	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	225	11	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	225	12	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	225	13	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	225	14	pre	pre	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	225	15	-	-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	225	16	existing	existing	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	225	17	baselines	baseline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	225	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	226	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	226	2	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	226	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	226	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	226	5	stan-	stan-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	6	dard	dard	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	7	phrase	phrase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	226	8	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	226	9	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	226	10	machine	machine	NN
work_lowlmiflivaeblgbyxkxrbt3nq	226	11	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	226	12	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	226	13	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	226	14	MOSES	MOSES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	15	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	226	16	Koehn	Koehn	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	17	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	18	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	19	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	226	21	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	226	22	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	226	23	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	226	24	GIZA++	GIZA++	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	25	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	226	26	Och	Och	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	27	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	226	28	Ney	Ney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	226	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	226	30	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	226	31	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	226	32	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	226	33	alignments	alignment	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	226	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	227	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	227	2	hold	hold	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	227	3	out	out	RP
work_lowlmiflivaeblgbyxkxrbt3nq	227	4	9	9	CD
work_lowlmiflivaeblgbyxkxrbt3nq	227	5	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	227	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	227	8	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	9	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	227	10	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	227	11	tuning	tuning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	12	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	227	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	227	14	MERT	MERT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	227	15	algo-	algo-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	16	rithm	rithm	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	227	18	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	227	19	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	227	20	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	227	21	Och	Och	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	227	22	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	227	23	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	227	24	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	227	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	228	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	228	2	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	228	3	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	228	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	228	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	228	6	implementation	implementation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	228	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	228	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	228	9	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	228	10	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	228	11	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	228	12	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	228	13	described	describe	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	228	14	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	228	15	Park	Park	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	228	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	228	17	Levy	Levy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	228	18	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	228	19	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	228	20	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	228	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	229	1	Like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	229	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	229	3	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	229	5	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	229	6	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	7	trains	train	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	229	8	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	229	9	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	10	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	229	11	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	229	12	EM	em	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	229	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	229	15	V	v	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	16	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	229	17	expectation	expectation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	18	semiring	semiring	NN
work_lowlmiflivaeblgbyxkxrbt3nq	229	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	230	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	230	2	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	3	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	230	5	crafted	craft	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	230	6	specifically	specifically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	230	7	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	230	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	230	9	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	10	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	11	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	230	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	230	14	so	so	RB
work_lowlmiflivaeblgbyxkxrbt3nq	230	15	differs	differ	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	230	16	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	230	17	Park	Park	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	230	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	230	19	Levy	Levy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	230	20	’s	’s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	230	21	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	230	22	several	several	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	230	23	ways	way	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	230	24	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	230	25	First	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	230	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	230	27	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	230	28	capture	capture	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	230	29	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	230	30	wider	wide	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	230	31	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	32	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	230	33	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	34	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	230	35	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	230	36	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	230	37	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	230	38	richer	rich	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	230	39	parameteri-	parameteri-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	230	40	zation	zation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	41	allowing	allow	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	230	42	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	230	43	more	more	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	230	44	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	230	45	options	option	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	230	46	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	231	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	231	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	231	3	result	result	NN
work_lowlmiflivaeblgbyxkxrbt3nq	231	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	231	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	231	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	231	7	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	231	8	6,718	6,718	CD
work_lowlmiflivaeblgbyxkxrbt3nq	231	9	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	231	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	231	11	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	231	12	more	more	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	231	13	than	than	IN
work_lowlmiflivaeblgbyxkxrbt3nq	231	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	231	15	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	231	16	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	231	17	’s	’s	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	231	18	187	187	CD
work_lowlmiflivaeblgbyxkxrbt3nq	231	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	232	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	232	2	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	232	3	corre-	corre-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	232	4	spond	spond	NN
work_lowlmiflivaeblgbyxkxrbt3nq	232	5	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	232	6	learned	learn	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	232	7	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	232	8	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	232	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	232	10	whereas	whereas	IN
work_lowlmiflivaeblgbyxkxrbt3nq	232	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	232	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	232	13	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	232	14	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	232	15	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	232	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	232	17	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	232	18	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	232	19	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	232	20	fixed	fix	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	232	21	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	232	22	uniform	uniform	NN
work_lowlmiflivaeblgbyxkxrbt3nq	232	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	233	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	233	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	233	3	capture	capture	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	233	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	233	5	larger	large	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	233	6	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	233	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	233	8	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	233	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	233	10	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	233	11	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	233	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	233	13	change	change	NN
work_lowlmiflivaeblgbyxkxrbt3nq	233	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	233	15	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	233	16	lemma	lemma	NN
work_lowlmiflivaeblgbyxkxrbt3nq	233	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	233	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	233	19	inflectional	inflectional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	233	20	overgeneralizations	overgeneralization	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	233	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	234	1	Finally	finally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	234	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	234	3	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	234	4	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	234	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	234	6	discriminative	discriminative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	234	7	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	234	8	step	step	NN
work_lowlmiflivaeblgbyxkxrbt3nq	234	9	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	234	10	model	model	VB
work_lowlmiflivaeblgbyxkxrbt3nq	234	11	long	long	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	234	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	234	13	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	234	14	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	234	15	dependencies	dependency	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	234	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	235	1	Although	although	IN
work_lowlmiflivaeblgbyxkxrbt3nq	235	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	235	3	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	235	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	235	5	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	235	6	originally	originally	RB
work_lowlmiflivaeblgbyxkxrbt3nq	235	7	geared	gear	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	235	8	towards	towards	IN
work_lowlmiflivaeblgbyxkxrbt3nq	235	9	fully	fully	RB
work_lowlmiflivaeblgbyxkxrbt3nq	235	10	unsupervised	unsupervise	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	235	11	train-	train-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	235	12	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	235	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	235	14	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	235	15	train	train	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	235	16	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	235	17	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	235	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	235	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	235	20	same	same	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	235	21	supervised	supervised	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	235	22	framework	framework	NN
work_lowlmiflivaeblgbyxkxrbt3nq	235	23	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	235	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	235	25	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	235	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	236	1	Evaluation	evaluation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	236	3	Performance	Performance	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	236	4	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	236	5	train	train	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	236	6	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	236	7	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	236	8	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	236	9	80	80	CD
work_lowlmiflivaeblgbyxkxrbt3nq	236	10	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	11	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	236	12	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	236	13	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	14	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	236	15	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	16	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	17	pairs	pair	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	236	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	236	19	test	test	VB
work_lowlmiflivaeblgbyxkxrbt3nq	236	20	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	236	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	236	22	remaining	remain	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	236	23	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	236	24	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	236	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	237	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	237	2	illustration	illustration	NN
work_lowlmiflivaeblgbyxkxrbt3nq	237	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	237	4	selected	select	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	237	5	output	output	NN
work_lowlmiflivaeblgbyxkxrbt3nq	237	6	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	237	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	237	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	237	9	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	237	10	shown	show	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	237	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	237	12	Table	Table	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	237	13	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	237	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	238	1	Predictions	prediction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	238	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	238	3	evaluated	evaluate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	238	4	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	5	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	6	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	7	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	8	Pap-	Pap-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	9	ineni	ineni	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	10	et	et	FW
work_lowlmiflivaeblgbyxkxrbt3nq	238	11	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	12	.	.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	238	14	2002	2002	CD
work_lowlmiflivaeblgbyxkxrbt3nq	238	15	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	238	17	Word	Word	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	18	Error	Error	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	19	Rate	rate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	20	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	21	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	22	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	238	24	de-	de-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	238	25	fined	fine	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	238	26	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	27	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	238	28	minimum	minimum	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	238	29	string	string	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	30	edit	edit	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	31	distance	distance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	32	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	33	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	34	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	238	35	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	238	36	between	between	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	37	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	38	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	238	39	predicted	predict	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	238	40	translations	translation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	238	41	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	238	42	di-	di-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	238	43	vided	vide	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	238	44	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	45	length	length	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	46	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	238	47	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	238	48	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	238	49	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	239	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	239	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	239	3	control	control	NN
work_lowlmiflivaeblgbyxkxrbt3nq	239	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	239	5	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	239	6	compare	compare	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	239	7	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	239	8	results	result	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	239	9	against	against	IN
work_lowlmiflivaeblgbyxkxrbt3nq	239	10	scores	score	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	239	11	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	239	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	239	13	uncor-	uncor-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	239	14	rected	recte	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	239	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	239	16	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	239	17	themselves	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	239	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	240	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	240	2	reported	report	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	240	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	240	4	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	240	5	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	240	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	240	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	240	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	240	9	achieves	achieve	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	240	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	240	11	best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	240	12	scores	score	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	240	13	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	240	14	both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	240	15	metrics	metric	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	240	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	241	1	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	241	2	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	241	3	increases	increase	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	241	4	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	5	50	50	CD
work_lowlmiflivaeblgbyxkxrbt3nq	241	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	7	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	241	8	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	241	9	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	10	62	62	CD
work_lowlmiflivaeblgbyxkxrbt3nq	241	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	241	12	while	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	13	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	241	14	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	241	15	reduced	reduce	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	241	16	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	17	.271	.271	CD
work_lowlmiflivaeblgbyxkxrbt3nq	241	18	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	241	19	.224	.224	CD
work_lowlmiflivaeblgbyxkxrbt3nq	241	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	242	1	Interestingly	interestingly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	242	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	242	3	MOSES	MOSES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	242	4	achieves	achieve	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	242	5	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	242	6	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	242	7	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	242	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	242	9	58	58	CD
work_lowlmiflivaeblgbyxkxrbt3nq	242	10	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	242	11	still	still	RB
work_lowlmiflivaeblgbyxkxrbt3nq	242	12	four	four	CD
work_lowlmiflivaeblgbyxkxrbt3nq	242	13	points	point	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	242	14	below	below	IN
work_lowlmiflivaeblgbyxkxrbt3nq	242	15	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	242	16	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	242	17	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	242	18	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	242	19	ac-	ac-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	242	20	tually	tually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	242	21	increases	increase	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	242	22	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	242	23	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	242	24	.449	.449	CD
work_lowlmiflivaeblgbyxkxrbt3nq	242	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	243	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	243	2	both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	243	3	metrics	metric	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	243	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	243	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	243	6	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	243	7	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	243	8	increases	increase	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	243	9	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	243	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	244	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	244	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	244	3	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	244	4	surprising	surprising	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	244	5	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	244	6	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	244	7	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	244	8	intended	intended	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	244	9	application	application	NN
work_lowlmiflivaeblgbyxkxrbt3nq	244	10	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	244	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	244	12	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	244	13	entirely	entirely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	244	14	different	different	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	244	15	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	244	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	245	1	Error	Error	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	245	2	Analysis	Analysis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	245	3	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	245	4	measured	measure	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	245	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	245	6	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	245	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	245	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	245	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	245	10	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	245	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	245	12	six	six	CD
work_lowlmiflivaeblgbyxkxrbt3nq	245	13	most	most	RBS
work_lowlmiflivaeblgbyxkxrbt3nq	245	14	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	245	15	categories	category	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	245	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	245	17	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	245	18	divergence	divergence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	245	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	245	20	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	245	21	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	245	22	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	245	23	various	various	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	245	24	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	245	25	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	245	26	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	245	27	overgeneralizations	overgeneralization	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	245	28	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	245	29	past	past	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	245	30	tense	tense	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	245	31	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	245	32	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	245	33	e.g.	e.g.	RB
work_lowlmiflivaeblgbyxkxrbt3nq	246	1	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	246	2	maked	make	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	246	3	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	246	4	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	246	5	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	246	6	made	make	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	246	7	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	246	8	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	246	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	247	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	247	2	first	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	247	3	iden-	iden-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	247	4	tified	tifie	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	247	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	6	parameters	parameter	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	247	7	associated	associate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	247	8	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	247	9	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	247	10	cate-	cate-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	11	gory	gory	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	247	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	247	14	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	247	15	counted	count	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	247	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	247	17	number	number	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	18	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	247	19	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	247	20	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	247	21	in-	in-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	247	22	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	247	23	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	24	firings	firing	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	247	25	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	247	26	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	247	27	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	247	28	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	247	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	248	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	248	2	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	248	3	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	248	4	indicates	indicate	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	248	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	248	6	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	248	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	248	8	performs	perform	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	248	9	reasonably	reasonably	RB
work_lowlmiflivaeblgbyxkxrbt3nq	248	10	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	248	11	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	248	12	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	248	13	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	248	14	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	248	15	verb	verb	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	248	16	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	248	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	248	18	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	248	19	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	248	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	248	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	248	22	overgeneralizations	overgeneralization	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	248	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	248	24	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	248	25	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	248	26	difficulty	difficulty	NN
work_lowlmiflivaeblgbyxkxrbt3nq	248	27	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	248	28	pronoun	pronoun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	248	29	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	248	30	auxiliary	auxiliary	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	248	31	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	248	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	249	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	249	2	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	249	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	249	4	hypothesizing	hypothesizing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	249	5	dropped	drop	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	249	6	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	249	7	burdens	burden	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	249	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	249	9	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	249	10	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	249	11	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	249	12	adding	add	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	249	13	additional	additional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	249	14	draws	draw	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	249	15	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	249	16	multinomial	multinomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	249	17	distributions	distribution	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	249	18	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	249	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	249	20	derivation	derivation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	249	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	250	1	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	250	2	pre-	pre-	VB
work_lowlmiflivaeblgbyxkxrbt3nq	250	3	132	132	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	4	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	5	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	6	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	7	reranking	reranke	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	250	8	62.12	62.12	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	9	.224	.224	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	10	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	11	reranking	reranke	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	250	12	60.86	60.86	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	13	.231	.231	.
work_lowlmiflivaeblgbyxkxrbt3nq	250	14	No	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	250	15	reranking	reranke	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	250	16	60.37	60.37	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	17	.233	.233	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	18	Moses	Moses	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	19	58.29	58.29	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	20	.449	.449	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	21	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	22	40.76	40.76	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	23	.318	.318	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	24	Child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	250	25	Sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	250	26	49.55	49.55	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	27	.271	.271	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	28	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	250	29	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	250	30	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	250	31	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	32	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	250	33	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	250	34	scores	score	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	250	35	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	251	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	251	2	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	3	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	251	4	perfor-	perfor-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	251	5	mance	mance	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	6	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	251	7	various	various	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	251	8	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	9	schemes	scheme	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	251	10	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	251	11	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	12	objec-	objec-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	13	tive	tive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	251	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	251	15	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	16	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	17	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	251	18	none	none	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	19	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	251	20	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	251	21	contrasted	contrast	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	251	22	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	251	23	Moses	Moses	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	24	MT	MT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	25	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	251	26	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	251	27	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	28	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	29	baselines	baseline	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	251	30	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	251	31	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	251	32	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	251	33	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	251	34	un-	un-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	251	35	corrected	corrected	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	251	36	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	251	37	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	251	38	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	252	1	Best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	252	2	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	252	3	under	under	IN
work_lowlmiflivaeblgbyxkxrbt3nq	252	4	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	252	5	metric	metric	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	252	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	252	7	shown	show	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	252	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	252	9	bold	bold	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	252	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	253	1	dict	dict	VB
work_lowlmiflivaeblgbyxkxrbt3nq	253	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	253	3	deletion	deletion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	253	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	253	5	either	either	CC
work_lowlmiflivaeblgbyxkxrbt3nq	253	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	253	7	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	253	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	253	9	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	253	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	253	11	reranker	reranker	NN
work_lowlmiflivaeblgbyxkxrbt3nq	253	12	must	must	MD
work_lowlmiflivaeblgbyxkxrbt3nq	253	13	strongly	strongly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	253	14	prefer	prefer	VB
work_lowlmiflivaeblgbyxkxrbt3nq	253	15	including	include	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	253	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	253	17	omit-	omit-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	253	18	ted	ted	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	253	19	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	253	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	254	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	254	2	syntax	syntax	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	254	4	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	254	5	noise	noise	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	7	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	254	8	achieve	achieve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	254	9	better	well	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	254	10	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	254	12	detecting	detect	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	254	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	254	14	correcting	correct	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	254	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	16	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	254	17	drops	drop	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	254	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	255	1	While	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	255	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	255	3	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	4	parameterization	parameterization	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	255	6	perfor-	perfor-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	7	mance	mance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	8	rely	rely	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	255	9	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	255	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	255	11	largely	largely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	255	12	constrained	constrained	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	255	13	nature	nature	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	255	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	16	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	255	17	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	255	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	255	19	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	255	20	observe	observe	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	255	21	some	some	DT
work_lowlmiflivaeblgbyxkxrbt3nq	255	22	instances	instance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	255	23	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	255	24	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	255	25	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	255	26	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	255	27	overly	overly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	255	28	restrictive	restrictive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	255	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	256	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	256	2	10	10	CD
work_lowlmiflivaeblgbyxkxrbt3nq	256	3	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	256	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	256	5	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	256	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	256	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	256	8	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	256	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	256	10	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	256	11	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	256	12	impossible	impossible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	256	13	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	256	14	recover	recover	VB
work_lowlmiflivaeblgbyxkxrbt3nq	256	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	256	16	exact	exact	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	256	17	gold	gold	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	256	18	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	256	19	standard	standard	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	256	20	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	256	21	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	256	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	257	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	257	2	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	257	3	fea-	fea-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	257	4	ture	ture	NN
work_lowlmiflivaeblgbyxkxrbt3nq	257	5	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	257	6	like	like	IN
work_lowlmiflivaeblgbyxkxrbt3nq	257	7	reordering	reordering	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	257	8	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	257	9	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	257	10	lemma	lemma	NN
work_lowlmiflivaeblgbyxkxrbt3nq	257	11	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	257	12	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	257	13	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	257	14	example	example	NN
work_lowlmiflivaeblgbyxkxrbt3nq	257	15	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	257	16	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	257	17	talk	talk	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	257	18	Mexican	Mexican	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	257	19	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	257	20	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	257	21	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	257	22	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	257	23	speak	speak	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	257	24	Spanish	Spanish	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	257	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	257	26	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	257	27	While	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	257	28	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	257	29	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	257	30	may	may	MD
work_lowlmiflivaeblgbyxkxrbt3nq	257	31	correct	correct	VB
work_lowlmiflivaeblgbyxkxrbt3nq	257	32	other	other	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	257	33	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	257	34	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	257	35	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	257	36	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	257	37	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	257	38	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	257	39	perfect	perfect	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	257	40	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	257	41	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	257	42	unattainable	unattainable	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	257	43	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	258	1	Sometimes	sometimes	RB
work_lowlmiflivaeblgbyxkxrbt3nq	258	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	258	3	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	258	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	258	5	produces	produce	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	258	6	appropriate	appropriate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	258	7	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	258	8	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	258	9	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	258	10	happenstance	happenstance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	258	11	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	258	12	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	258	13	conform	conform	VB
work_lowlmiflivaeblgbyxkxrbt3nq	258	14	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	258	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	258	16	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	258	17	’	’	POS
work_lowlmiflivaeblgbyxkxrbt3nq	258	18	decision	decision	NN
work_lowlmiflivaeblgbyxkxrbt3nq	258	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	259	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	259	2	example	example	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	259	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	259	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	6	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	259	7	row	row	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	259	9	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	10	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	259	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	259	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	13	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	14	corrects	correct	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	259	15	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	259	16	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	17	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	259	18	have	have	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	259	19	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	20	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	259	21	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	259	22	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	259	23	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	259	24	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	25	one	one	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	26	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	259	27	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	28	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	259	29	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	259	30	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	259	31	instead	instead	RB
work_lowlmiflivaeblgbyxkxrbt3nq	259	32	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	259	33	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	34	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	259	35	verbose	verbose	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	259	36	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	37	chosen	choose	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	259	38	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	259	39	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	40	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	259	41	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	259	42	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	259	43	Does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	259	44	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	259	45	one	one	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	259	46	have	have	VB
work_lowlmiflivaeblgbyxkxrbt3nq	259	47	water	water	NN
work_lowlmiflivaeblgbyxkxrbt3nq	259	48	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	259	49	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	259	50	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	259	51	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	260	1	Similarly	similarly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	260	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	260	3	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	260	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	260	5	sometimes	sometimes	RB
work_lowlmiflivaeblgbyxkxrbt3nq	260	6	produces	produce	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	260	7	corrections	correction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	260	8	which	which	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	260	9	seem	seem	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	260	10	appro-	appro-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	260	11	priate	priate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	260	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	260	13	isolation	isolation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	260	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	260	15	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	260	16	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	260	17	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	260	18	preserve	preserve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	260	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	260	20	meaning	meaning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	260	21	implied	imply	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	260	22	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	260	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	260	24	larger	large	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	260	25	conversational	conversational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	260	26	context	context	NN
work_lowlmiflivaeblgbyxkxrbt3nq	260	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	261	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	261	2	ex-	ex-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	261	3	ample	ample	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	261	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	261	6	row	row	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	7	three	three	CD
work_lowlmiflivaeblgbyxkxrbt3nq	261	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	261	9	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	10	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	261	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	261	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	261	13	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	14	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	261	15	Want	want	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	261	16	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	261	17	read	read	VB
work_lowlmiflivaeblgbyxkxrbt3nq	261	18	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	261	19	book	book	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	261	21	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	261	22	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	261	23	recognized	recognize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	261	24	both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	261	25	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	261	26	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	261	27	hu-	hu-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	261	28	man	man	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	29	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	261	30	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	261	31	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	261	32	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	33	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	261	34	requiring	require	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	261	35	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	261	36	pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	37	noun	noun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	38	subject	subject	NN
work_lowlmiflivaeblgbyxkxrbt3nq	261	39	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	262	1	Unlike	unlike	IN
work_lowlmiflivaeblgbyxkxrbt3nq	262	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	3	annotators	annotator	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	262	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	262	5	however	however	RB
work_lowlmiflivaeblgbyxkxrbt3nq	262	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	262	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	8	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	9	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	262	10	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	11	knowledge	knowledge	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	12	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	262	13	conversational	conversational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	262	14	context	context	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	262	16	so	so	CC
work_lowlmiflivaeblgbyxkxrbt3nq	262	17	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	262	18	chooses	choose	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	262	19	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	20	highest	high	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	262	21	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	22	pronoun	pronoun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	23	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	262	24	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	262	25	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	26	case	case	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	27	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	262	28	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	262	29	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	262	30	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	262	31	instead	instead	RB
work_lowlmiflivaeblgbyxkxrbt3nq	262	32	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	262	33	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	262	34	contextually	contextually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	262	35	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	262	36	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	262	37	I.	i.	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	38	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	262	39	Error	Error	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	262	40	Type	Type	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	262	41	Count	Count	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	262	42	F1	F1	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	262	43	P	p	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	44	R	r	NN
work_lowlmiflivaeblgbyxkxrbt3nq	262	45	Be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	262	46	Deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	262	47	63	63	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	48	.84	.84	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	49	.84	.84	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	50	.84	.84	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	51	Pronoun	Pronoun	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	262	52	Deletions	Deletions	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	262	53	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	54	.15	.15	CD
work_lowlmiflivaeblgbyxkxrbt3nq	262	55	.38	.38	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	262	56	.1	.1	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	262	57	Aux	aux	UH
work_lowlmiflivaeblgbyxkxrbt3nq	262	58	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	263	1	Deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	263	2	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	263	3	.21	.21	CD
work_lowlmiflivaeblgbyxkxrbt3nq	263	4	.44	.44	CD
work_lowlmiflivaeblgbyxkxrbt3nq	263	5	.13	.13	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	263	6	Prep	Prep	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	263	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	264	1	Deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	264	2	26	26	CD
work_lowlmiflivaeblgbyxkxrbt3nq	264	3	.65	.65	CD
work_lowlmiflivaeblgbyxkxrbt3nq	264	4	.82	.82	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	264	5	.54	.54	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	264	6	Det	Det	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	264	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	265	1	Deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	265	2	22	22	CD
work_lowlmiflivaeblgbyxkxrbt3nq	265	3	.48	.48	CD
work_lowlmiflivaeblgbyxkxrbt3nq	265	4	.73	.73	CD
work_lowlmiflivaeblgbyxkxrbt3nq	265	5	.36	.36	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	265	6	Overgen	Overgen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	265	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	266	1	Past	past	IN
work_lowlmiflivaeblgbyxkxrbt3nq	266	2	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	3	.92	.92	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	4	1.0	1.0	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	5	.86	.86	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	6	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	7	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	8	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	266	9	Frequency	frequency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	266	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	266	12	six	six	CD
work_lowlmiflivaeblgbyxkxrbt3nq	266	13	most	most	RBS
work_lowlmiflivaeblgbyxkxrbt3nq	266	14	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	266	15	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	16	types	type	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	266	17	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	266	18	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	19	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	266	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	266	21	along	along	IN
work_lowlmiflivaeblgbyxkxrbt3nq	266	22	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	266	23	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	266	24	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	25	’s	’s	,
work_lowlmiflivaeblgbyxkxrbt3nq	266	26	corresponding	corresponding	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	266	27	F-	F-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	266	28	measure	measure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	266	30	precision	precision	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	31	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	266	32	recall	recall	NN
work_lowlmiflivaeblgbyxkxrbt3nq	266	33	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	267	1	All	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	267	2	counts	count	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	267	3	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	267	4	±.12	±.12	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	267	5	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	267	6	p	p	NN
work_lowlmiflivaeblgbyxkxrbt3nq	267	7	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	267	8	.05	.05	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	267	9	under	under	IN
work_lowlmiflivaeblgbyxkxrbt3nq	267	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	267	11	binomial	binomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	267	12	normal	normal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	267	13	approximation	approximation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	267	14	inter-	inter-	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	267	15	val	val	NN
work_lowlmiflivaeblgbyxkxrbt3nq	267	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	268	1	0	0	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	2	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	3	10	10	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	4	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	5	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	6	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	7	30	30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	8	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	9	40	40	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	10	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	11	50	50	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	12	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	13	60	60	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	14	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	15	70	70	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	16	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	17	80	80	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	18	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	19	90	90	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	20	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	21	100	100	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	22	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	23	40	40	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	24	45	45	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	25	50	50	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	26	55	55	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	27	60	60	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	28	65	65	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	29	.20	.20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	30	.22	.22	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	31	.24	.24	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	32	.26	.26	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	268	33	.28	.28	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	34	.30	.30	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	35	.32	.32	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	36	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	37	Train	Train	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	268	38	Data	Data	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	268	39	B	B	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	268	40	L	l	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	41	E	e	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	42	U	u	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	43	W	w	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	44	E	e	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	45	R	r	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	46	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	47	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	268	48	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	268	49	Performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	50	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	268	51	limited	limited	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	268	52	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	268	53	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	268	54	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	269	1	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	269	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	269	3	drawn	draw	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	269	4	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	269	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	269	6	dashed	dash	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	269	7	line	line	NN
work_lowlmiflivaeblgbyxkxrbt3nq	269	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	269	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	269	10	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	269	11	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	269	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	269	13	solid	solid	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	269	14	line	line	NN
work_lowlmiflivaeblgbyxkxrbt3nq	269	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	270	1	Learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	270	2	Curves	curve	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	270	3	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	270	4	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	270	5	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	270	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	270	7	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	270	8	see	see	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	270	9	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	270	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	270	11	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	270	12	curves	curve	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	270	13	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	270	14	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	270	15	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	270	16	initially	initially	RB
work_lowlmiflivaeblgbyxkxrbt3nq	270	17	rise	rise	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	270	18	sharply	sharply	RB
work_lowlmiflivaeblgbyxkxrbt3nq	270	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	270	20	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	270	21	remain	remain	VB
work_lowlmiflivaeblgbyxkxrbt3nq	270	22	relatively	relatively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	270	23	flat	flat	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	270	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	271	1	Using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	271	2	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	271	3	10	10	CD
work_lowlmiflivaeblgbyxkxrbt3nq	271	4	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	271	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	271	6	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	271	7	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	271	8	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	271	9	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	271	10	80	80	CD
work_lowlmiflivaeblgbyxkxrbt3nq	271	11	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	271	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	271	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	271	14	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	271	15	increase	increase	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	271	16	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	271	17	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	271	18	44	44	CD
work_lowlmiflivaeblgbyxkxrbt3nq	271	19	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	271	20	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	271	21	just	just	RB
work_lowlmiflivaeblgbyxkxrbt3nq	271	22	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	271	23	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	271	24	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	271	25	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	271	26	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	271	27	almost	almost	RB
work_lowlmiflivaeblgbyxkxrbt3nq	271	28	61	61	CD
work_lowlmiflivaeblgbyxkxrbt3nq	271	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	272	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	272	2	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	272	3	reach	reach	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	272	4	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	272	5	reported	report	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	272	6	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	272	7	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	272	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	272	9	62	62	CD
work_lowlmiflivaeblgbyxkxrbt3nq	272	10	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	272	11	adding	add	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	272	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	272	13	final	final	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	272	14	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	272	15	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	272	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	272	17	training	train	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	272	18	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	272	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	273	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	273	2	result	result	NN
work_lowlmiflivaeblgbyxkxrbt3nq	273	3	emphasizes	emphasize	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	273	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	273	5	specificity	specificity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	273	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	273	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	273	8	parameteri-	parameteri-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	273	9	zation	zation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	273	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	274	1	Because	because	IN
work_lowlmiflivaeblgbyxkxrbt3nq	274	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	274	3	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	4	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	274	5	so	so	RB
work_lowlmiflivaeblgbyxkxrbt3nq	274	6	tailored	tailor	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	274	7	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	274	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	274	9	child-	child-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	10	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	11	scenario	scenario	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	274	13	only	only	RB
work_lowlmiflivaeblgbyxkxrbt3nq	274	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	274	15	few	few	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	274	16	examples	example	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	274	17	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	274	18	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	274	19	er-	er-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	274	20	ror	ror	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	21	type	type	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	22	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	274	23	needed	need	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	274	24	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	274	25	find	find	VB
work_lowlmiflivaeblgbyxkxrbt3nq	274	26	good	good	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	274	27	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	274	28	values	value	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	274	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	275	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	275	2	suspect	suspect	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	275	3	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	275	4	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	275	5	annotated	annotate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	275	6	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	275	7	would	would	MD
work_lowlmiflivaeblgbyxkxrbt3nq	275	8	lead	lead	VB
work_lowlmiflivaeblgbyxkxrbt3nq	275	9	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	275	10	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	275	11	continued	continued	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	275	12	but	but	CC
work_lowlmiflivaeblgbyxkxrbt3nq	275	13	slow	slow	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	275	14	increase	increase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	275	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	275	16	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	275	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	276	1	Training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	276	3	Testing	testing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	4	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	276	5	Children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	276	6	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	276	7	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	276	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	276	9	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	10	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	276	11	investigate	investigate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	276	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	276	13	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	14	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	276	15	lan-	lan-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	276	16	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	17	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	18	follows	follow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	276	19	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	276	20	similar	similar	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	276	21	path	path	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	22	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	276	23	chil-	chil-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	276	24	dren	dren	NN
work_lowlmiflivaeblgbyxkxrbt3nq	276	25	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	276	26	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	276	27	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	276	28	1973	1973	CD
work_lowlmiflivaeblgbyxkxrbt3nq	276	29	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	276	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	277	1	To	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	277	2	test	test	VB
work_lowlmiflivaeblgbyxkxrbt3nq	277	3	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	277	4	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	277	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	277	6	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	277	7	train	train	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	277	8	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	277	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	277	10	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	277	11	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	277	12	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	277	13	excluding	exclude	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	277	14	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	277	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	277	16	who	who	WP
work_lowlmiflivaeblgbyxkxrbt3nq	277	17	alone	alone	RB
work_lowlmiflivaeblgbyxkxrbt3nq	277	18	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	277	19	responsible	responsible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	277	20	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	277	21	21	21	CD
work_lowlmiflivaeblgbyxkxrbt3nq	277	22	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	277	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	277	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	277	25	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	277	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	278	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	278	2	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	278	3	test	test	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	278	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	278	5	learned	learn	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	278	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	278	7	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	278	8	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	278	9	separated	separate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	278	10	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	278	11	133	133	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	12	Trained	train	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	278	13	on	on	RB
work_lowlmiflivaeblgbyxkxrbt3nq	278	14	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	278	15	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	278	16	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	278	17	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	278	18	72.58	72.58	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	19	.226	.226	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	20	All	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	278	21	Others	other	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	278	22	69.83	69.83	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	23	.186	.186	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	24	Uncorrected	uncorrecte	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	278	25	45.54	45.54	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	26	.278	.278	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	27	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	278	28	5	5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	278	29	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	278	30	Performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	278	31	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	278	32	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	278	33	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	278	34	sentences	sentence	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	278	35	training	train	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	278	36	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	278	37	other	other	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	278	38	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	278	39	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	278	40	versus	versus	IN
work_lowlmiflivaeblgbyxkxrbt3nq	278	41	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	278	42	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	278	43	himself	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	278	44	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	279	1	Best	good	JJS
work_lowlmiflivaeblgbyxkxrbt3nq	279	2	perfor-	perfor-	IN
work_lowlmiflivaeblgbyxkxrbt3nq	279	3	mance	mance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	279	4	under	under	IN
work_lowlmiflivaeblgbyxkxrbt3nq	279	5	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	279	6	metric	metric	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	279	7	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	279	8	shown	show	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	279	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	279	10	bold	bold	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	279	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	280	1	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	280	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	281	1	These	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	281	2	results	result	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	281	3	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	281	4	contrasted	contrast	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	281	5	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	281	6	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	281	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	281	8	8-fold	8-fold	CD
work_lowlmiflivaeblgbyxkxrbt3nq	281	9	cross	cross	NN
work_lowlmiflivaeblgbyxkxrbt3nq	281	10	validation	validation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	281	11	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	281	12	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	281	13	testing	testing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	281	14	solely	solely	RB
work_lowlmiflivaeblgbyxkxrbt3nq	281	15	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	281	16	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	281	17	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	281	18	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	281	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	282	1	Performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	282	2	statistics	statistic	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	282	3	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	282	4	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	282	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	282	6	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	282	7	5	5	CD
work_lowlmiflivaeblgbyxkxrbt3nq	282	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	283	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	283	2	first	first	RB
work_lowlmiflivaeblgbyxkxrbt3nq	283	3	note	note	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	283	4	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	283	5	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	283	6	trained	train	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	283	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	283	8	both	both	DT
work_lowlmiflivaeblgbyxkxrbt3nq	283	9	scenar-	scenar-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	283	10	ios	ios	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	283	11	lead	lead	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	283	12	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	283	13	large	large	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	283	14	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	283	15	reductions	reduction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	283	16	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	283	17	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	283	18	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	283	19	sen-	sen-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	283	20	tences	tence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	283	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	284	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	284	2	provides	provide	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	284	3	evidence	evidence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	284	4	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	284	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	284	6	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	284	7	cap-	cap-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	284	8	tures	ture	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	284	9	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	284	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	284	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	284	12	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	284	13	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	284	14	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	284	15	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	284	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	284	17	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	284	18	patterns	pattern	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	284	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	285	1	Although	although	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	2	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	285	3	exclusively	exclusively	RB
work_lowlmiflivaeblgbyxkxrbt3nq	285	4	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	5	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	285	6	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	285	7	lead	lead	VB
work_lowlmiflivaeblgbyxkxrbt3nq	285	8	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	9	increased	increase	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	285	10	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	285	11	score	score	NN
work_lowlmiflivaeblgbyxkxrbt3nq	285	12	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	285	13	72.58	72.58	CD
work_lowlmiflivaeblgbyxkxrbt3nq	285	14	vs	vs	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	15	69.83	69.83	CD
work_lowlmiflivaeblgbyxkxrbt3nq	285	16	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	285	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	285	18	WER	WER	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	285	19	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	285	20	minimized	minimize	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	285	21	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	285	22	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	285	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	285	24	larger	large	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	285	25	volume	volume	NN
work_lowlmiflivaeblgbyxkxrbt3nq	285	26	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	27	train-	train-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	285	28	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	285	29	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	285	30	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	31	other	other	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	285	32	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	285	33	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	285	34	.186	.186	CD
work_lowlmiflivaeblgbyxkxrbt3nq	285	35	vs	vs	IN
work_lowlmiflivaeblgbyxkxrbt3nq	285	36	.226	.226	CD
work_lowlmiflivaeblgbyxkxrbt3nq	285	37	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	285	38	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	286	1	Taken	take	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	286	2	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	286	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	286	4	whole	whole	NN
work_lowlmiflivaeblgbyxkxrbt3nq	286	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	286	6	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	286	7	results	result	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	286	8	suggest	suggest	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	286	9	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	286	10	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	286	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	286	12	testing	testing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	286	13	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	286	14	separate	separate	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	286	15	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	286	16	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	286	17	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	286	18	degrade	degrade	VB
work_lowlmiflivaeblgbyxkxrbt3nq	286	19	perfor-	perfor-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	286	20	mance	mance	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	286	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	287	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	287	2	finding	finding	NN
work_lowlmiflivaeblgbyxkxrbt3nq	287	3	supports	support	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	287	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	287	5	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	287	6	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	287	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	287	8	shared	share	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	287	9	developmental	developmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	287	10	paths	path	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	287	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	288	1	Plotting	plot	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	288	2	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	3	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	4	Errors	Errors	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	288	5	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	6	Time	Time	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	7	Af-	Af-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	8	ter	ter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	9	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	10	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	11	annotated	annotate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	288	12	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	288	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	288	14	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	288	15	predict	predict	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	288	16	diver-	diver-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	288	17	gences	gence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	288	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	19	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	288	20	available	available	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	288	21	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	288	22	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	288	24	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	288	25	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	26	Roger	Roger	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	27	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	28	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	288	29	1973	1973	CD
work_lowlmiflivaeblgbyxkxrbt3nq	288	30	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	31	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	288	32	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	288	34	Eve	Eve	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	35	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	288	36	Sarah	Sarah	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	37	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	288	38	as	as	RB
work_lowlmiflivaeblgbyxkxrbt3nq	288	39	well	well	RB
work_lowlmiflivaeblgbyxkxrbt3nq	288	40	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	41	Abe	Abe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	42	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	288	43	Kuczaj	Kuczaj	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	288	44	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	288	45	1977	1977	CD
work_lowlmiflivaeblgbyxkxrbt3nq	288	46	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	288	47	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	288	48	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	288	49	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	50	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	51	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	288	52	sep-	sep-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	53	arate	arate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	54	study	study	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	55	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	288	56	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	288	57	similar	similar	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	288	58	age	age	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	59	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	288	60	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	288	61	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	289	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	289	2	plot	plot	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	289	3	each	each	DT
work_lowlmiflivaeblgbyxkxrbt3nq	289	4	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	289	5	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	289	6	per	per	IN
work_lowlmiflivaeblgbyxkxrbt3nq	289	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	289	8	utterance	utterance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	289	9	frequency	frequency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	289	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	289	11	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	289	12	omis-	omis-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	289	13	sions	sion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	289	14	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	289	15	Figure	Figure	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	289	16	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	289	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	290	1	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	290	2	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	290	3	evaluate	evaluate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	290	4	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	290	5	65,000	65,000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	290	6	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	290	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	290	8	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	9	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	290	10	no	no	DT
work_lowlmiflivaeblgbyxkxrbt3nq	290	11	impact	impact	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	12	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	290	13	preposi-	preposi-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	290	14	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	15	drop	drop	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	16	prediction	prediction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	290	18	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	290	19	skip	skip	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	290	20	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	290	21	reranking	reranking	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	22	step	step	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	23	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	290	24	save	save	VB
work_lowlmiflivaeblgbyxkxrbt3nq	290	25	computation	computation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	290	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	291	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	2	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	3	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	291	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	291	5	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	291	6	see	see	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	291	7	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	8	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	291	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	291	10	Sarah	Sarah	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	291	11	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	291	12	prepo-	prepo-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	13	sition	sition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	14	drops	drop	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	291	15	spike	spike	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	291	16	early	early	RB
work_lowlmiflivaeblgbyxkxrbt3nq	291	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	291	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	291	19	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	291	20	gradually	gradually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	291	21	decrease	decrease	VB
work_lowlmiflivaeblgbyxkxrbt3nq	291	22	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	23	frequency	frequency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	24	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	25	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	291	26	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	27	use	use	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	291	28	moves	move	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	291	29	towards	towards	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	30	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	291	31	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	291	32	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	291	33	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	291	34	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	292	1	Although	although	IN
work_lowlmiflivaeblgbyxkxrbt3nq	292	2	Eve	Eve	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	292	3	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	292	4	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	292	5	covers	cover	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	292	6	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	292	7	ear-	ear-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	292	8	lier	li	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	292	9	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	10	period	period	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	292	12	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	292	13	see	see	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	292	14	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	292	15	her	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	292	16	pattern	pattern	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	17	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	292	18	prepo-	prepo-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	292	19	sition	sition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	20	drops	drop	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	292	21	shows	show	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	292	22	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	292	23	similar	similar	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	292	24	spike	spike	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	292	25	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	292	26	gradual	gradual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	292	27	de-	de-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	28	crease	crease	NN
work_lowlmiflivaeblgbyxkxrbt3nq	292	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	293	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	293	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	293	3	consistent	consistent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	293	4	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	293	5	Eve	Eve	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	293	6	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	293	7	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	293	8	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	293	9	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	293	10	precocity	precocity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	293	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	294	1	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	294	2	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	294	3	conclusion	conclusion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	294	4	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	294	5	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	294	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	294	7	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	294	8	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	294	9	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	294	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	294	11	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	294	12	three	three	CD
work_lowlmiflivaeblgbyxkxrbt3nq	294	13	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	294	14	advanced	advance	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	294	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	294	16	similar	similar	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	294	17	stages	stage	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	294	18	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	294	19	different	different	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	294	20	times	time	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	294	21	—	—	:
work_lowlmiflivaeblgbyxkxrbt3nq	294	22	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	294	23	consistent	consistent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	294	24	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	294	25	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	294	26	predictions	prediction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	294	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	295	1	However	however	RB
work_lowlmiflivaeblgbyxkxrbt3nq	295	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	295	3	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	295	4	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	295	5	examine	examine	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	295	6	18	18	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	7	23	23	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	8	28	28	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	9	33	33	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	10	38	38	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	11	43	43	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	12	48	48	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	13	53	53	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	14	58	58	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	15	0	0	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	16	0.02	0.02	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	17	0.04	0.04	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	18	0.06	0.06	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	19	0.08	0.08	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	20	0.1	0.1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	21	0.12	0.12	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	22	0.14	0.14	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	23	Adam	Adam	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	24	Eve	Eve	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	25	Sarah	Sarah	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	26	Abe	Abe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	27	Age	Age	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	28	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	295	29	Months	Months	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	30	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	295	31	P	p	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	32	e	e	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	33	r-	r-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	295	34	U	U	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	35	tt	tt	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	36	e	e	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	37	ra	ra	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	38	n	n	CC
work_lowlmiflivaeblgbyxkxrbt3nq	295	39	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	40	e	e	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	41	F	F	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	42	re	re	IN
work_lowlmiflivaeblgbyxkxrbt3nq	295	43	q	q	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	44	u	u	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	45	e	e	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	46	n	n	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	47	c	c	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	48	y	y	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	49	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	50	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	51	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	295	52	Automatically	automatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	295	53	detected	detect	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	295	54	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	55	omissions	omission	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	295	56	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	295	57	un	un	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	295	58	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	295	59	annotated	annotate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	295	60	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	295	61	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	295	62	four	four	CD
work_lowlmiflivaeblgbyxkxrbt3nq	295	63	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	295	64	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	295	65	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	295	66	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	296	1	Assuming	assume	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	296	2	perfect	perfect	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	296	3	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	296	4	predictions	prediction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	296	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	296	6	frequencies	frequency	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	296	7	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	296	8	±.002	±.002	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	296	9	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	296	10	p	p	NN
work_lowlmiflivaeblgbyxkxrbt3nq	296	11	=	=	SYM
work_lowlmiflivaeblgbyxkxrbt3nq	296	12	.05	.05	NFP
work_lowlmiflivaeblgbyxkxrbt3nq	296	13	under	under	IN
work_lowlmiflivaeblgbyxkxrbt3nq	296	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	296	15	binomial	binomial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	296	16	normal	normal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	296	17	approxima-	approxima-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	296	18	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	296	19	interval	interval	NN
work_lowlmiflivaeblgbyxkxrbt3nq	296	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	297	1	Prediction	prediction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	297	2	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	297	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	297	4	given	give	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	297	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	297	6	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	297	7	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	297	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	298	1	Abe	Abe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	298	2	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	298	3	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	298	4	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	298	5	observe	observe	VB
work_lowlmiflivaeblgbyxkxrbt3nq	298	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	298	7	same	same	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	298	8	pattern.7	pattern.7	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	9	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	298	10	points	point	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	298	11	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	12	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	298	13	degree	degree	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	15	variance	variance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	16	across	across	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	17	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	298	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	298	19	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	298	20	suggests	suggest	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	298	21	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	298	22	use	use	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	298	25	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	26	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	27	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	298	28	tool	tool	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	29	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	30	further	further	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	298	31	empirical	empirical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	298	32	refinement	refinement	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	33	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	298	34	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	35	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	298	36	hy-	hy-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	298	37	potheses	pothese	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	298	38	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	299	1	Discussion	Discussion	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	299	2	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	299	3	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	4	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	5	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	6	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	299	7	de-	de-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	299	8	signed	sign	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	299	9	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	299	10	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	299	11	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	299	12	constrained	constrain	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	299	13	than	than	IN
work_lowlmiflivaeblgbyxkxrbt3nq	299	14	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	299	15	full	full	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	299	16	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	299	17	scale	scale	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	18	MT	MT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	299	19	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	299	21	focusing	focus	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	299	22	parameter	parameter	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	23	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	299	24	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	299	25	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	299	26	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	299	27	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	299	28	known	know	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	299	29	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	299	30	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	299	31	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	299	32	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	299	33	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	34	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	299	35	learn-	learn-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	299	36	ers	ers	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	299	37	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	300	1	Reorderings	reordering	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	300	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	300	3	prohibited	prohibit	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	300	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	300	5	lexical	lexical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	300	6	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	300	7	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	300	8	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	300	9	limited	limit	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	300	10	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	300	11	inflectional	inflectional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	300	12	changes	change	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	300	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	300	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	300	15	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	300	16	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	300	17	restricted	restrict	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	300	18	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	300	19	function	function	VB
work_lowlmiflivaeblgbyxkxrbt3nq	300	20	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	300	21	categories	category	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	300	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	301	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	301	2	highly	highly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	301	3	re-	re-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	301	4	stricting	stricte	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	301	5	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	301	6	hypothesis	hypothesis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	7	space	space	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	301	9	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	301	10	provide	provide	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	301	11	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	301	12	induc-	induc-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	13	tive	tive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	301	14	bias	bias	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	15	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	301	16	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	301	17	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	18	that	that	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	301	19	matches	match	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	301	20	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	301	21	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	22	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	23	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	24	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	301	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	302	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	302	2	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	302	3	particularly	particularly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	302	4	important	important	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	302	5	since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	302	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	302	7	size	size	NN
work_lowlmiflivaeblgbyxkxrbt3nq	302	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	302	9	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	302	10	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	302	11	set	set	NN
work_lowlmiflivaeblgbyxkxrbt3nq	302	12	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	302	13	much	much	RB
work_lowlmiflivaeblgbyxkxrbt3nq	302	14	smaller	small	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	302	15	than	than	IN
work_lowlmiflivaeblgbyxkxrbt3nq	302	16	that	that	DT
work_lowlmiflivaeblgbyxkxrbt3nq	302	17	usually	usually	RB
work_lowlmiflivaeblgbyxkxrbt3nq	302	18	used	use	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	302	19	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	302	20	MT	MT	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	302	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	303	1	Indeed	indeed	RB
work_lowlmiflivaeblgbyxkxrbt3nq	303	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	303	3	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	303	4	Figure	figure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	303	5	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	303	6	shows	show	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	303	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	303	8	very	very	RB
work_lowlmiflivaeblgbyxkxrbt3nq	303	9	little	little	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	303	10	data	data	NN
work_lowlmiflivaeblgbyxkxrbt3nq	303	11	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	303	12	needed	need	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	303	13	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	303	14	achieve	achieve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	303	15	good	good	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	303	16	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	303	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	304	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	304	2	contrast	contrast	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	304	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	304	5	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	304	6	baseline	baseline	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	7	suffers	suffer	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	304	8	because	because	IN
work_lowlmiflivaeblgbyxkxrbt3nq	304	9	its	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	304	10	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	304	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	12	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	304	13	too	too	RB
work_lowlmiflivaeblgbyxkxrbt3nq	304	14	restricted	restrict	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	304	15	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	304	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	304	17	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	18	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	304	19	transcribed	transcribed	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	304	20	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	21	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	304	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	305	1	As	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	305	2	shown	show	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	305	3	above	above	RB
work_lowlmiflivaeblgbyxkxrbt3nq	305	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	305	5	Table	table	NN
work_lowlmiflivaeblgbyxkxrbt3nq	305	6	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	305	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	305	8	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	305	9	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	305	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	305	11	function	function	NN
work_lowlmiflivaeblgbyxkxrbt3nq	305	12	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	305	13	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	305	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	305	15	most	most	RBS
work_lowlmiflivaeblgbyxkxrbt3nq	305	16	frequent	frequent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	305	17	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	305	18	types	type	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	305	19	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	305	20	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	305	21	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	305	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	306	1	Since	since	IN
work_lowlmiflivaeblgbyxkxrbt3nq	306	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	306	3	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	306	4	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	5	does	do	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	306	6	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	306	7	capture	capture	VB
work_lowlmiflivaeblgbyxkxrbt3nq	306	8	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	9	deletions	deletion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	306	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	306	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	306	12	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	306	13	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	306	14	more	more	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	306	15	restricted	restricted	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	306	16	notion	notion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	17	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	306	18	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	19	swaps	swap	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	306	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	306	21	88	88	CD
work_lowlmiflivaeblgbyxkxrbt3nq	306	22	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	23	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	306	24	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	25	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	306	26	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	306	27	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	306	28	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	29	corpus	corpus	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	306	30	can	can	MD
work_lowlmiflivaeblgbyxkxrbt3nq	306	31	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	306	32	be	be	VB
work_lowlmiflivaeblgbyxkxrbt3nq	306	33	translated	translate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	306	34	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	306	35	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	306	36	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	37	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	306	38	versions	version	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	306	39	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	307	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	307	2	result	result	NN
work_lowlmiflivaeblgbyxkxrbt3nq	307	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	307	4	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	307	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	307	6	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	307	7	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	307	8	tends	tend	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	307	9	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	307	10	rely	rely	VB
work_lowlmiflivaeblgbyxkxrbt3nq	307	11	too	too	RB
work_lowlmiflivaeblgbyxkxrbt3nq	307	12	heavily	heavily	RB
work_lowlmiflivaeblgbyxkxrbt3nq	307	13	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	307	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	307	15	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	307	16	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	307	17	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	307	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	308	1	For	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	308	2	example	example	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	308	4	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	308	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	308	6	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	7	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	308	8	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	308	9	com-	com-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	308	10	7Though	7though	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	11	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	308	12	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	308	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	308	14	course	course	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	15	possible	possible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	308	16	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	308	17	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	308	18	similar	similar	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	308	19	spike	spike	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	20	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	308	21	drop	drop	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	22	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	308	23	off	off	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	24	occurred	occur	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	308	25	earlier	early	RBR
work_lowlmiflivaeblgbyxkxrbt3nq	308	26	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	308	27	Abe	Abe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	308	28	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	308	29	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	308	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	309	1	134	134	CD
work_lowlmiflivaeblgbyxkxrbt3nq	309	2	ing	e	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	309	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	309	4	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	309	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	309	6	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	309	7	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	309	8	ESL	ESL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	309	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	309	10	improves	improve	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	309	11	n	n	NN
work_lowlmiflivaeblgbyxkxrbt3nq	309	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	309	13	gram	gram	NN
work_lowlmiflivaeblgbyxkxrbt3nq	309	14	prob-	prob-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	309	15	ability	ability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	309	16	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	309	17	producing	produce	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	309	18	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	309	19	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	309	20	came	come	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	309	21	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	309	22	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	309	23	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	309	24	instead	instead	RB
work_lowlmiflivaeblgbyxkxrbt3nq	309	25	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	309	26	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	309	27	correct	correct	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	309	28	“	"	``
work_lowlmiflivaeblgbyxkxrbt3nq	309	29	I	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	309	30	am	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	309	31	coming	come	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	309	32	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	309	33	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	309	34	”	"	''
work_lowlmiflivaeblgbyxkxrbt3nq	309	35	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	310	1	This	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	310	2	increases	increase	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	310	3	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	310	4	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	310	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	310	6	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	310	7	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	310	8	itself	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	310	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	311	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	311	2	addition	addition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	311	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	311	5	domain	domain	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	311	7	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	311	8	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	311	9	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	311	11	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	311	12	approach	approach	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	13	has	have	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	311	14	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	311	15	advantage	advantage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	16	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	311	17	long-	long-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	311	18	range	range	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	19	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	311	20	information	information	NN
work_lowlmiflivaeblgbyxkxrbt3nq	311	21	encoded	encode	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	311	22	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	311	23	reranking	reranke	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	311	24	features	feature	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	311	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	312	1	Although	although	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	312	3	perceptron	perceptron	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	312	4	algorithm	algorithm	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	312	5	places	place	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	312	6	high	high	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	312	7	weight	weight	NN
work_lowlmiflivaeblgbyxkxrbt3nq	312	8	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	9	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	312	10	generative	generative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	312	11	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	312	12	probability	probability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	312	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	312	14	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	312	15	alters	alter	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	312	16	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	312	17	predictions	prediction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	312	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	19	17	17	CD
work_lowlmiflivaeblgbyxkxrbt3nq	312	20	out	out	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	21	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	22	201	201	CD
work_lowlmiflivaeblgbyxkxrbt3nq	312	23	test	test	NN
work_lowlmiflivaeblgbyxkxrbt3nq	312	24	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	312	25	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	312	26	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	312	27	all	all	DT
work_lowlmiflivaeblgbyxkxrbt3nq	312	28	cases	case	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	312	29	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	312	30	improvement	improvement	NN
work_lowlmiflivaeblgbyxkxrbt3nq	312	31	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	313	1	Three	three	CD
work_lowlmiflivaeblgbyxkxrbt3nq	313	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	313	3	these	these	DT
work_lowlmiflivaeblgbyxkxrbt3nq	313	4	rerank-	rerank-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	313	5	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	6	changes	change	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	313	7	add	add	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	313	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	313	9	noun	noun	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	10	subject	subject	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	313	12	five	five	CD
work_lowlmiflivaeblgbyxkxrbt3nq	313	13	enforce	enforce	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	14	ques-	ques-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	313	15	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	16	structure	structure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	313	18	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	313	19	nine	nine	CD
work_lowlmiflivaeblgbyxkxrbt3nq	313	20	add	add	VB
work_lowlmiflivaeblgbyxkxrbt3nq	313	21	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	313	22	main	main	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	313	23	verb	verb	NN
work_lowlmiflivaeblgbyxkxrbt3nq	313	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	314	1	7	7	CD
work_lowlmiflivaeblgbyxkxrbt3nq	314	2	Conclusion	Conclusion	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	314	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	314	4	Future	Future	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	314	5	Work	Work	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	314	6	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	314	7	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	314	8	paper	paper	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	9	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	314	10	introduce	introduce	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	314	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	314	12	corpus	corpus	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	314	14	divergent	divergent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	314	15	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	16	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	314	17	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	314	18	corresponding	correspond	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	314	19	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	20	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	314	21	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	314	22	en-	en-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	314	23	abling	able	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	314	24	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	314	25	systematic	systematic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	314	26	computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	314	27	modeling	modeling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	28	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	314	29	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	30	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	31	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	314	32	relating	relate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	314	33	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	314	34	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	314	35	adult	adult	VB
work_lowlmiflivaeblgbyxkxrbt3nq	314	36	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	314	37	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	315	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	315	2	propose	propose	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	315	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	315	4	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	5	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	315	6	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	315	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	315	8	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	9	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	10	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	11	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	315	12	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	315	13	means	means	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	14	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	315	15	investigate	investigate	VB
work_lowlmiflivaeblgbyxkxrbt3nq	315	16	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	17	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	18	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	315	20	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	315	21	pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	22	vide	vide	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	315	23	an	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	315	24	initial	initial	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	315	25	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	26	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	315	27	this	this	DT
work_lowlmiflivaeblgbyxkxrbt3nq	315	28	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	315	29	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	316	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	316	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	316	3	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	316	4	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	316	5	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	316	6	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	316	7	noisy	noisy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	316	8	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	316	9	channel	channel	NN
work_lowlmiflivaeblgbyxkxrbt3nq	316	10	assump-	assump-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	316	11	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	316	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	316	13	allowing	allow	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	316	14	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	316	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	316	16	deletion	deletion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	316	17	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	316	18	corruption	corruption	NN
work_lowlmiflivaeblgbyxkxrbt3nq	316	19	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	316	20	in-	in-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	316	21	dividual	dividual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	316	22	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	316	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	316	24	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	316	25	is	be	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	316	26	trained	train	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	316	27	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	316	28	FST	FST	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	316	29	techniques	technique	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	316	30	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	317	1	Despite	despite	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	2	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	317	3	debatable	debatable	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	4	cognitive	cognitive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	5	plausibility	plausibility	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	7	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	317	8	setup	setup	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	317	10	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	317	11	results	result	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	317	12	demonstrate	demonstrate	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	317	13	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	14	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	317	15	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	16	cap-	cap-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	17	tures	ture	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	317	18	many	many	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	19	standard	standard	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	20	divergences	divergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	317	21	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	317	22	reduces	reduce	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	317	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	317	24	average	average	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	25	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	26	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	27	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	28	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	317	29	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	30	approximately	approximately	RB
work_lowlmiflivaeblgbyxkxrbt3nq	317	31	20	20	CD
work_lowlmiflivaeblgbyxkxrbt3nq	317	32	%	%	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	33	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	317	34	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	35	high	high	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	36	performance	performance	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	37	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	317	38	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	317	39	frequently	frequently	RB
work_lowlmiflivaeblgbyxkxrbt3nq	317	40	occurring	occur	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	317	41	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	317	42	types	type	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	317	43	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	318	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	318	2	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	3	allows	allow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	318	4	us	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	318	5	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	318	6	chart	chart	VB
work_lowlmiflivaeblgbyxkxrbt3nq	318	7	aspects	aspect	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	318	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	318	9	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	10	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	11	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	318	12	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	318	14	without	without	IN
work_lowlmiflivaeblgbyxkxrbt3nq	318	15	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	318	16	need	need	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	17	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	318	18	addi-	addi-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	318	19	tional	tional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	318	20	human	human	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	318	21	annotation	annotation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	318	22	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	319	1	Our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	319	2	experiments	experiment	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	319	3	show	show	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	319	4	that	that	IN
work_lowlmiflivaeblgbyxkxrbt3nq	319	5	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	319	6	share	share	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	319	7	common	common	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	319	8	developmental	developmental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	319	9	stages	stage	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	319	10	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	319	11	lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	12	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	13	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	319	15	while	while	IN
work_lowlmiflivaeblgbyxkxrbt3nq	319	16	pointing	point	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	319	17	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	319	18	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	19	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	319	20	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	319	21	sub-	sub-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	319	22	tleties	tletie	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	319	23	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	319	24	preposition	preposition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	25	use	use	NN
work_lowlmiflivaeblgbyxkxrbt3nq	319	26	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	320	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	320	2	future	future	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	320	3	work	work	NN
work_lowlmiflivaeblgbyxkxrbt3nq	320	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	320	5	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	320	6	intend	intend	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	320	7	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	320	8	dynamically	dynamically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	320	9	model	model	VB
work_lowlmiflivaeblgbyxkxrbt3nq	320	10	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	320	11	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	320	12	ability	ability	NN
work_lowlmiflivaeblgbyxkxrbt3nq	320	13	as	as	IN
work_lowlmiflivaeblgbyxkxrbt3nq	320	14	it	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	320	15	grows	grow	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	320	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	320	17	shifts	shift	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	320	18	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	320	19	re-	re-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	320	20	sponse	sponse	NN
work_lowlmiflivaeblgbyxkxrbt3nq	320	21	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	320	22	internal	internal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	320	23	processes	process	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	320	24	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	320	25	external	external	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	320	26	stimuli	stimulus	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	320	27	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	321	1	We	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	321	2	also	also	RB
work_lowlmiflivaeblgbyxkxrbt3nq	321	3	plan	plan	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	321	4	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	321	5	develop	develop	VB
work_lowlmiflivaeblgbyxkxrbt3nq	321	6	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	321	7	train	train	VB
work_lowlmiflivaeblgbyxkxrbt3nq	321	8	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	321	9	specializ-	specializ-	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	321	10	ing	ing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	321	11	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	321	12	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	321	13	detection	detection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	321	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	321	15	specific	specific	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	321	16	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	321	17	categories	category	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	321	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	322	1	By	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	2	explicitly	explicitly	RB
work_lowlmiflivaeblgbyxkxrbt3nq	322	3	shifting	shift	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	322	4	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	322	5	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	6	’s	’s	,
work_lowlmiflivaeblgbyxkxrbt3nq	322	7	objective	objective	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	8	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	9	child-	child-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	322	10	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	11	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	12	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	322	14	detection	detection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	15	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	16	some	some	DT
work_lowlmiflivaeblgbyxkxrbt3nq	322	17	particular	particular	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	322	18	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	19	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	322	20	we	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	322	21	hope	hope	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	322	22	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	322	23	improve	improve	VB
work_lowlmiflivaeblgbyxkxrbt3nq	322	24	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	322	25	analysis	analysis	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	26	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	27	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	28	di-	di-	CC
work_lowlmiflivaeblgbyxkxrbt3nq	322	29	vergences	vergence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	322	30	over	over	IN
work_lowlmiflivaeblgbyxkxrbt3nq	322	31	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	322	32	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	323	1	Acknowledgments	acknowledgment	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	323	2	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	323	3	authors	author	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	323	4	thank	thank	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	323	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	323	6	reviewers	reviewer	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	323	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	323	8	acknowledge	acknowledge	NN
work_lowlmiflivaeblgbyxkxrbt3nq	323	9	support	support	NN
work_lowlmiflivaeblgbyxkxrbt3nq	323	10	by	by	IN
work_lowlmiflivaeblgbyxkxrbt3nq	323	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	323	12	NSF	NSF	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	323	13	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	323	14	grant	grant	NN
work_lowlmiflivaeblgbyxkxrbt3nq	323	15	IIS-1116676	IIS-1116676	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	323	16	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	323	17	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	323	18	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	323	19	re-	re-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	323	20	search	search	NN
work_lowlmiflivaeblgbyxkxrbt3nq	323	21	gift	gift	NN
work_lowlmiflivaeblgbyxkxrbt3nq	323	22	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	323	23	Google	Google	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	323	24	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	324	1	Any	any	DT
work_lowlmiflivaeblgbyxkxrbt3nq	324	2	opinions	opinion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	324	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	324	4	findings	finding	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	324	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	324	6	or	or	CC
work_lowlmiflivaeblgbyxkxrbt3nq	324	7	conclusions	conclusion	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	324	8	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	324	9	those	those	DT
work_lowlmiflivaeblgbyxkxrbt3nq	324	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	324	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	324	12	authors	author	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	324	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	324	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	324	15	do	do	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	324	16	not	not	RB
work_lowlmiflivaeblgbyxkxrbt3nq	324	17	nec-	nec-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	324	18	essarily	essarily	RB
work_lowlmiflivaeblgbyxkxrbt3nq	324	19	reflect	reflect	VB
work_lowlmiflivaeblgbyxkxrbt3nq	324	20	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	324	21	views	view	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	324	22	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	324	23	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	324	24	NSF	NSF	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	324	25	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	325	1	References	reference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	325	2	A.	a.	NN
work_lowlmiflivaeblgbyxkxrbt3nq	325	3	Alishahi	Alishahi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	325	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	326	1	2010	2010	CD
work_lowlmiflivaeblgbyxkxrbt3nq	326	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	327	1	Computational	computational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	327	2	modeling	modeling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	327	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	327	4	human	human	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	327	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	327	6	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	327	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	328	1	Synthesis	Synthesis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	328	2	Lectures	Lectures	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	328	3	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	328	4	Human	Human	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	328	5	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	328	6	Technologies	Technologies	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	328	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	328	8	3(1):1–107	3(1):1–107	CD
work_lowlmiflivaeblgbyxkxrbt3nq	328	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	329	1	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	2	Allauzen	Allauzen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	329	4	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	5	Riley	Riley	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	329	7	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	8	Schalkwyk	Schalkwyk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	329	10	W.	W.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	11	Skut	Skut	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	329	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	329	14	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	15	Mohri	Mohri	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	329	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	330	1	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	330	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	331	1	OpenFst	openfst	NN
work_lowlmiflivaeblgbyxkxrbt3nq	331	2	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	331	3	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	331	4	general	general	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	331	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	331	6	efficient	efficient	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	331	7	weighted	weighted	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	331	8	finite	finite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	331	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	331	10	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	331	11	transducer	transducer	NN
work_lowlmiflivaeblgbyxkxrbt3nq	331	12	library	library	NN
work_lowlmiflivaeblgbyxkxrbt3nq	331	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	332	1	Implementa-	Implementa-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	332	2	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	332	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	332	4	Application	Application	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	332	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	332	6	Automata	Automata	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	332	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	332	8	pages	page	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	332	9	11–23	11–23	CD
work_lowlmiflivaeblgbyxkxrbt3nq	332	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	333	1	R.H.	R.H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	2	Baayen	Baayen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	333	4	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	5	Piepenbrock	Piepenbrock	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	333	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	333	8	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	9	Gulikers	Gulikers	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	333	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	334	1	1996	1996	CD
work_lowlmiflivaeblgbyxkxrbt3nq	334	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	335	1	CELEX2	CELEX2	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	335	2	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	335	3	CD	cd	NN
work_lowlmiflivaeblgbyxkxrbt3nq	335	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	335	5	ROM	rom	NN
work_lowlmiflivaeblgbyxkxrbt3nq	335	6	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	335	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	336	1	Linguistic	Linguistic	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	336	2	Data	Data	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	336	3	Consortium	Consortium	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	336	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	337	1	E.	E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	2	Bates	Bates	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	337	4	I.	I.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	5	Bretherton	Bretherton	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	337	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	337	8	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	9	Snyder	Snyder	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	337	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	338	1	1988	1988	CD
work_lowlmiflivaeblgbyxkxrbt3nq	338	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	339	1	From	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	339	2	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	339	3	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	339	4	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	339	5	grammar	grammar	VB
work_lowlmiflivaeblgbyxkxrbt3nq	339	6	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	339	7	Individual	individual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	339	8	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	339	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	339	10	disso-	disso-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	339	11	ciable	ciable	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	339	12	mechanisms	mechanism	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	339	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	340	1	Cambridge	Cambridge	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	340	2	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	340	3	Press	Press	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	340	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	341	1	D.C.	D.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	341	2	Bellinger	Bellinger	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	341	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	341	4	J.B.	J.B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	341	5	Gleason	Gleason	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	341	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	342	1	1982	1982	CD
work_lowlmiflivaeblgbyxkxrbt3nq	342	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	343	1	Sex	sex	NN
work_lowlmiflivaeblgbyxkxrbt3nq	343	2	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	343	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	343	4	parental	parental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	343	5	directives	directive	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	343	6	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	343	7	young	young	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	343	8	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	343	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	344	1	Sex	sex	NN
work_lowlmiflivaeblgbyxkxrbt3nq	344	2	Roles	Roles	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	344	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	344	4	8(11):1123–1139	8(11):1123–1139	CD
work_lowlmiflivaeblgbyxkxrbt3nq	344	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	345	1	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	345	2	Bliss	Bliss	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	345	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	346	1	1988	1988	CD
work_lowlmiflivaeblgbyxkxrbt3nq	346	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	347	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	347	2	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	347	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	347	4	modals	modal	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	347	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	348	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	348	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	348	3	Applied	Applied	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	348	4	Developmental	Developmental	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	348	5	Psychology	Psychology	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	348	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	348	7	9:253–261	9:253–261	CD
work_lowlmiflivaeblgbyxkxrbt3nq	348	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	349	1	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	2	Bloom	Bloom	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	349	4	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	5	Hood	Hood	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	349	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	349	8	P.	P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	9	Lightbown	Lightbown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	349	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	350	1	1974	1974	CD
work_lowlmiflivaeblgbyxkxrbt3nq	350	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	351	1	Imitation	imitation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	351	2	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	351	3	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	351	4	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	351	5	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	351	6	If	if	IN
work_lowlmiflivaeblgbyxkxrbt3nq	351	7	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	351	8	when	when	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	351	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	351	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	351	11	why	why	WRB
work_lowlmiflivaeblgbyxkxrbt3nq	351	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	352	1	Cognitive	Cognitive	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	352	2	Psychology	Psychology	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	352	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	352	4	6(3):380–420	6(3):380–420	CD
work_lowlmiflivaeblgbyxkxrbt3nq	352	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	353	1	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	2	Bloom	Bloom	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	353	4	P.	P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	5	Lightbown	Lightbown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	353	7	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	8	Hood	Hood	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	353	10	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	11	Bowerman	Bowerman	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	353	13	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	14	Maratsos	Maratsos	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	353	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	353	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	353	17	M.P.	M.P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	354	1	Maratsos	Maratsos	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	354	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	355	1	1975	1975	CD
work_lowlmiflivaeblgbyxkxrbt3nq	355	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	356	1	Structure	structure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	356	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	356	3	variation	variation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	356	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	356	5	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	356	6	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	356	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	357	1	Monographs	Monographs	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	357	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	357	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	357	4	Soci-	soci-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	357	5	ety	ety	NN
work_lowlmiflivaeblgbyxkxrbt3nq	357	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	357	7	Research	Research	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	357	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	357	9	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	357	10	Development	Development	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	357	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	357	12	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	357	13	1–97	1–97	CD
work_lowlmiflivaeblgbyxkxrbt3nq	357	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	358	1	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	358	2	Bloom	Bloom	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	358	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	359	1	1973	1973	CD
work_lowlmiflivaeblgbyxkxrbt3nq	359	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	360	1	One	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	360	2	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	360	3	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	360	4	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	360	5	time	time	NN
work_lowlmiflivaeblgbyxkxrbt3nq	360	6	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	360	7	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	360	8	use	use	NN
work_lowlmiflivaeblgbyxkxrbt3nq	360	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	360	10	single	single	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	360	11	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	360	12	utterances	utterance	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	360	13	before	before	IN
work_lowlmiflivaeblgbyxkxrbt3nq	360	14	syntax	syntax	NN
work_lowlmiflivaeblgbyxkxrbt3nq	360	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	361	1	Mouton	Mouton	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	361	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	362	1	P.	P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	362	2	Bloom	Bloom	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	362	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	363	1	1990	1990	CD
work_lowlmiflivaeblgbyxkxrbt3nq	363	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	364	1	Subjectless	subjectless	NN
work_lowlmiflivaeblgbyxkxrbt3nq	364	2	sentences	sentence	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	364	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	364	4	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	364	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	364	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	365	1	Linguistic	Linguistic	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	365	2	Inquiry	Inquiry	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	365	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	365	4	21(4):491–504	21(4):491–504	CD
work_lowlmiflivaeblgbyxkxrbt3nq	365	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	366	1	J.N.	J.N.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	367	1	Bohannon	Bohannon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	367	2	III	III	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	367	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	367	4	A.L.	A.L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	368	1	Marquis	Marquis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	368	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	369	1	1977	1977	CD
work_lowlmiflivaeblgbyxkxrbt3nq	369	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	370	1	Chil-	Chil-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	370	2	dren	dren	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	370	3	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	370	4	control	control	NN
work_lowlmiflivaeblgbyxkxrbt3nq	370	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	370	6	adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	370	7	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	370	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	371	1	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	371	2	Development	Development	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	371	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	371	4	48(3):1002–1008	48(3):1002–1008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	371	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	372	1	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	372	2	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	372	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	373	1	1973	1973	CD
work_lowlmiflivaeblgbyxkxrbt3nq	373	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	374	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	374	2	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	374	3	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	374	4	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	374	5	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	374	6	early	early	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	374	7	stages	stage	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	374	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	375	1	Harvard	Harvard	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	375	2	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	375	3	Press	Press	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	375	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	376	1	V.	V.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	376	2	Carlson	Carlson	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	376	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	376	4	Luden	Luden	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	376	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	377	1	1979	1979	CD
work_lowlmiflivaeblgbyxkxrbt3nq	377	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	378	1	Causal	causal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	378	2	understanding	understanding	NN
work_lowlmiflivaeblgbyxkxrbt3nq	378	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	378	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	378	5	10-month	10-month	CD
work_lowlmiflivaeblgbyxkxrbt3nq	378	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	378	7	old	old	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	378	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	379	1	Ph.D.	Ph.D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	379	2	thesis	thesis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	379	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	379	4	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	379	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	379	6	Colorado	Colorado	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	379	7	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	379	8	Boulder	Boulder	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	379	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	380	1	E.C.	E.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	381	1	Carterette	Carterette	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	381	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	381	3	M.H.	M.H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	382	1	Jones	Jones	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	382	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	383	1	1974	1974	CD
work_lowlmiflivaeblgbyxkxrbt3nq	383	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	384	1	Informal	informal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	384	2	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	384	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	384	4	Alphabetic	alphabetic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	384	5	&	&	CC
work_lowlmiflivaeblgbyxkxrbt3nq	384	6	phonemic	phonemic	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	384	7	texts	text	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	384	8	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	384	9	statistical	statistical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	384	10	analyses	analysis	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	384	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	384	12	tables	table	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	384	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	385	1	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	385	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	385	3	California	California	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	385	4	Press	Press	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	385	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	386	1	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	386	2	Chodorow	Chodorow	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	386	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	386	4	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	386	5	Leacock	Leacock	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	386	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	387	1	2000	2000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	387	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	388	1	An	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	388	2	unsupervised	unsupervised	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	388	3	method	method	NN
work_lowlmiflivaeblgbyxkxrbt3nq	388	4	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	388	5	detecting	detect	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	388	6	grammatical	grammatical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	388	7	errors	error	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	388	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	389	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	389	2	Proceed-	Proceed-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	3	ings	ing	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	389	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	389	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	389	6	North	north	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	389	7	American	American	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	8	Chapter	Chapter	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	389	10	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	389	11	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	12	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	389	13	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	14	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	389	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	389	16	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	389	17	140–147	140–147	CD
work_lowlmiflivaeblgbyxkxrbt3nq	389	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	390	1	135	135	CD
work_lowlmiflivaeblgbyxkxrbt3nq	390	2	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	390	3	Collins	Collins	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	390	4	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	390	5	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	390	6	Roark	Roark	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	390	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	391	1	2004	2004	CD
work_lowlmiflivaeblgbyxkxrbt3nq	391	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	392	1	Incremental	incremental	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	392	2	parsing	parse	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	392	3	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	392	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	392	5	perceptron	perceptron	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	392	6	algorithm	algorithm	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	392	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	393	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	393	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	393	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	393	5	Asso-	Asso-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	6	ciation	ciation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	393	7	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	393	8	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	9	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	393	11	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	393	12	111–118	111–118	CD
work_lowlmiflivaeblgbyxkxrbt3nq	393	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	393	14	Barcelona	Barcelona	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	393	16	Spain	Spain	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	393	18	July	July	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	393	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	394	1	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	2	Connor	Connor	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	394	4	Y.	Y.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	5	Gertner	Gertner	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	394	7	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	8	Fisher	Fisher	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	394	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	394	11	D.	D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	12	Roth	Roth	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	394	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	395	1	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	395	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	396	1	Baby	baby	NN
work_lowlmiflivaeblgbyxkxrbt3nq	396	2	SRL	SRL	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	396	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	396	4	Modeling	model	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	396	5	early	early	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	396	6	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	396	7	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	396	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	397	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	397	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	397	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	397	5	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	397	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	8	Nat-	Nat-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	9	ural	ural	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	397	10	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	11	Learning	Learning	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	397	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	397	13	pages	page	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	397	14	81–88	81–88	CD
work_lowlmiflivaeblgbyxkxrbt3nq	397	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	398	1	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	398	2	Dale	Dale	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	398	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	398	4	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	398	5	Kilgarriff	Kilgarriff	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	398	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	399	1	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	399	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	400	1	Helping	help	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	400	2	our	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	400	3	own	own	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	400	4	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	400	5	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	400	6	HOO	HOO	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	400	7	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	400	8	pilot	pilot	NN
work_lowlmiflivaeblgbyxkxrbt3nq	400	9	shared	share	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	400	10	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	400	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	401	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	401	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	401	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	401	5	Eu-	Eu-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	6	ropean	ropean	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	401	7	Workshop	Workshop	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	8	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	401	9	Natural	Natural	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	10	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	11	Generation	Generation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	401	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	401	13	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	401	14	242–249	242–249	CD
work_lowlmiflivaeblgbyxkxrbt3nq	401	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	402	1	M.C.	M.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	403	1	de	de	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	403	2	Marneffe	Marneffe	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	403	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	403	4	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	403	5	MacCartney	MacCartney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	403	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	403	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	403	8	C.D.	C.D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	404	1	Manning	manning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	404	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	405	1	2006	2006	CD
work_lowlmiflivaeblgbyxkxrbt3nq	405	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	406	1	Generating	generate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	406	2	typed	type	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	406	3	dependency	dependency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	406	4	parses	parse	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	406	5	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	406	6	phrase	phrase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	406	7	structure	structure	NN
work_lowlmiflivaeblgbyxkxrbt3nq	406	8	parses	parse	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	406	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	407	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	407	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	407	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	407	4	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	407	5	In-	In-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	407	6	ternational	ternational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	407	7	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	407	8	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	407	9	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	407	10	Resources	Resources	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	407	11	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	407	12	Evaluation	Evaluation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	407	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	407	14	volume	volume	NN
work_lowlmiflivaeblgbyxkxrbt3nq	407	15	6	6	CD
work_lowlmiflivaeblgbyxkxrbt3nq	407	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	407	17	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	407	18	449–454	449–454	CD
work_lowlmiflivaeblgbyxkxrbt3nq	407	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	408	1	M.J.	M.J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	408	2	Demetras	Demetras	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	408	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	408	4	K.N.	K.N.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	409	1	Post	Post	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	409	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	409	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	409	4	C.E.	C.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	410	1	Snow	snow	NN
work_lowlmiflivaeblgbyxkxrbt3nq	410	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	411	1	1986	1986	CD
work_lowlmiflivaeblgbyxkxrbt3nq	411	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	412	1	Feed-	feed-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	412	2	back	back	RB
work_lowlmiflivaeblgbyxkxrbt3nq	412	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	412	4	first	first	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	412	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	412	6	learners	learner	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	412	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	412	8	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	412	9	role	role	NN
work_lowlmiflivaeblgbyxkxrbt3nq	412	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	412	11	repetitions	repetition	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	412	12	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	412	13	clarification	clarification	NN
work_lowlmiflivaeblgbyxkxrbt3nq	412	14	questions	question	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	412	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	413	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	413	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	413	3	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	413	4	Lan-	Lan-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	413	5	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	413	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	413	7	13(2):275–292	13(2):275–292	CD
work_lowlmiflivaeblgbyxkxrbt3nq	413	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	414	1	M.J.	M.J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	414	2	Demetras	Demetras	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	414	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	415	1	1989	1989	CD
work_lowlmiflivaeblgbyxkxrbt3nq	415	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	416	1	Working	work	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	416	2	parents	parent	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	416	3	’	’	POS
work_lowlmiflivaeblgbyxkxrbt3nq	416	4	conversational	conversational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	416	5	responses	response	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	416	6	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	416	7	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	416	8	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	416	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	416	10	year	year	NN
work_lowlmiflivaeblgbyxkxrbt3nq	416	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	416	12	old	old	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	416	13	sons	son	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	416	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	417	1	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	2	Dreyer	Dreyer	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	417	4	J.R.	J.R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	5	Smith	Smith	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	417	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	417	8	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	9	Eisner	Eisner	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	417	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	418	1	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	418	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	419	1	Latent-	Latent-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	419	2	variable	variable	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	419	3	modeling	modeling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	419	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	419	5	string	string	NN
work_lowlmiflivaeblgbyxkxrbt3nq	419	6	transductions	transduction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	419	7	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	419	8	finite-	finite-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	419	9	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	419	10	methods	method	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	419	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	420	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	420	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	420	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	420	5	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	420	7	Empirical	Empirical	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	8	Methods	Methods	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	420	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	420	10	Natural	Natural	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	11	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	12	Processing	Processing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	420	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	420	14	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	420	15	1080–1089	1080–1089	CD
work_lowlmiflivaeblgbyxkxrbt3nq	420	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	421	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	421	2	Echihabi	Echihabi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	421	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	421	4	D.	D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	421	5	Marcu	Marcu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	421	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	422	1	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	422	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	423	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	423	2	noisy	noisy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	423	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	423	4	channel	channel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	423	5	ap-	ap-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	423	6	proach	proach	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	423	7	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	423	8	question	question	VB
work_lowlmiflivaeblgbyxkxrbt3nq	423	9	answering	answering	NN
work_lowlmiflivaeblgbyxkxrbt3nq	423	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	424	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	424	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	424	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	424	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	424	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	424	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	424	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	424	8	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	424	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	424	10	pages	page	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	424	11	16	16	CD
work_lowlmiflivaeblgbyxkxrbt3nq	424	12	–	–	:
work_lowlmiflivaeblgbyxkxrbt3nq	424	13	23	23	CD
work_lowlmiflivaeblgbyxkxrbt3nq	424	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	425	1	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	425	2	Eisner	Eisner	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	425	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	426	1	2001	2001	CD
work_lowlmiflivaeblgbyxkxrbt3nq	426	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	427	1	Expectation	expectation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	427	2	semirings	semiring	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	427	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	427	4	Flexible	flexible	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	427	5	EM	em	NN
work_lowlmiflivaeblgbyxkxrbt3nq	427	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	427	7	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	427	8	finite	finite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	427	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	427	10	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	427	11	transducers	transducer	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	427	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	428	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	428	2	Proceedings	proceeding	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	428	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	428	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	428	5	ESSLLI	ESSLLI	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	428	6	workshop	workshop	NN
work_lowlmiflivaeblgbyxkxrbt3nq	428	7	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	428	8	finite	finite	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	428	9	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	428	10	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	428	11	methods	method	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	428	12	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	428	13	NLP	NLP	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	428	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	429	1	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	429	2	Gamon	Gamon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	429	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	430	1	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	430	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	431	1	High	high	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	431	2	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	431	3	order	order	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	4	sequence	sequence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	5	modeling	model	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	431	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	431	7	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	8	learner	learner	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	9	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	10	detection	detection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	431	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	432	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	432	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	432	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	432	5	Workshop	Workshop	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	432	7	Innovative	innovative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	432	8	Use	Use	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	9	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	432	10	NLP	NLP	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	11	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	432	12	Building	Building	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	13	Ed-	Ed-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	432	14	ucational	ucational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	432	15	Applications	Applications	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	432	16	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	432	17	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	432	18	180–189	180–189	CD
work_lowlmiflivaeblgbyxkxrbt3nq	432	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	433	1	L.C.G.	L.C.G.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	434	1	Haggerty	haggerty	RB
work_lowlmiflivaeblgbyxkxrbt3nq	434	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	435	1	1930	1930	CD
work_lowlmiflivaeblgbyxkxrbt3nq	435	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	436	1	What	what	WDT
work_lowlmiflivaeblgbyxkxrbt3nq	436	2	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	436	3	two	two	CD
work_lowlmiflivaeblgbyxkxrbt3nq	436	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	436	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	436	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	436	7	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	436	8	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	436	9	half	half	NN
work_lowlmiflivaeblgbyxkxrbt3nq	436	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	436	11	year-	year-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	436	12	old	old	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	436	13	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	436	14	said	say	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	436	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	436	16	one	one	CD
work_lowlmiflivaeblgbyxkxrbt3nq	436	17	day	day	NN
work_lowlmiflivaeblgbyxkxrbt3nq	436	18	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	437	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	437	2	Pedagogical	Pedagogical	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	437	3	Seminary	Seminary	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	437	4	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	437	5	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	437	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	437	7	Genetic	Genetic	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	437	8	Psychology	Psychology	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	437	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	437	10	37(1):75–101	37(1):75–101	CD
work_lowlmiflivaeblgbyxkxrbt3nq	437	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	438	1	W.S.	W.S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	439	1	Hall	Hall	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	439	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	439	3	W.C.	W.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	439	4	Tirre	Tirre	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	439	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	439	6	A.L.	A.L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	440	1	Brown	Brown	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	440	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	440	3	J.C.	J.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	440	4	Campoine	Campoine	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	440	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	440	6	P.F.	P.F.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	1	Nardulli	Nardulli	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	441	3	HO	HO	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	4	Abdulrahman	Abdulrahman	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	441	6	MA	MA	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	7	Sozen	Sozen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	441	9	W.C.	W.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	10	Schno-	Schno-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	11	brich	brich	NN
work_lowlmiflivaeblgbyxkxrbt3nq	441	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	441	13	H.	H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	14	Cecen	Cecen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	441	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	441	16	J.G.	J.G.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	442	1	Barnitz	Barnitz	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	442	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	442	3	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	442	4	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	442	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	443	1	1979	1979	CD
work_lowlmiflivaeblgbyxkxrbt3nq	443	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	444	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	444	2	communicative	communicative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	444	3	environment	environment	NN
work_lowlmiflivaeblgbyxkxrbt3nq	444	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	444	5	young	young	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	444	6	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	444	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	444	8	Social	social	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	444	9	class	class	NN
work_lowlmiflivaeblgbyxkxrbt3nq	444	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	444	11	ethnic	ethnic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	444	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	444	13	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	444	14	situational	situational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	444	15	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	444	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	445	1	Bulletin	bulletin	NN
work_lowlmiflivaeblgbyxkxrbt3nq	445	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	445	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	445	4	Center	Center	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	445	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	445	6	Children	Children	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	445	7	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	445	8	Books	Books	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	445	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	445	10	32:08	32:08	CD
work_lowlmiflivaeblgbyxkxrbt3nq	445	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	446	1	W.S.	W.S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	447	1	Hall	Hall	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	447	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	447	3	W.E.	W.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	448	1	Nagy	Nagy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	448	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	448	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	448	4	R.L.	R.L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	449	1	Linn	Linn	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	449	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	450	1	1980	1980	CD
work_lowlmiflivaeblgbyxkxrbt3nq	450	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	451	1	Spoken	speak	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	451	2	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	451	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	451	4	Effects	effect	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	451	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	451	6	situation	situation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	451	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	451	8	social	social	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	451	9	group	group	NN
work_lowlmiflivaeblgbyxkxrbt3nq	451	10	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	451	11	oral	oral	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	451	12	word	word	NN
work_lowlmiflivaeblgbyxkxrbt3nq	451	13	usage	usage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	451	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	451	15	frequency	frequency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	451	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	452	1	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	452	3	Illinois	Illinois	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	4	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	452	5	Urbana	Urbana	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	452	7	Champaign	Champaign	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	452	9	Center	Center	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	10	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	452	11	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	452	12	Study	Study	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	452	14	Reading	Reading	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	452	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	453	1	W.S.	W.S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	454	1	Hall	Hall	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	454	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	454	3	W.E.	W.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	455	1	Nagy	Nagy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	455	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	455	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	455	4	G.	G.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	455	5	Nottenburg	Nottenburg	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	455	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	456	1	1981	1981	CD
work_lowlmiflivaeblgbyxkxrbt3nq	456	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	457	1	Sit-	Sit-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	457	2	uational	uational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	457	3	variation	variation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	457	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	457	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	457	6	use	use	NN
work_lowlmiflivaeblgbyxkxrbt3nq	457	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	457	8	internal	internal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	457	9	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	457	10	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	457	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	458	1	Technical	technical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	458	2	report	report	NN
work_lowlmiflivaeblgbyxkxrbt3nq	458	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	458	4	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	458	6	Illinois	Illinois	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	7	at	at	IN
work_lowlmiflivaeblgbyxkxrbt3nq	458	8	Urbana-	Urbana-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	9	Champaign	Champaign	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	458	11	Center	Center	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	12	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	458	13	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	458	14	Study	Study	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	15	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	458	16	Reading	Reading	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	458	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	459	1	H.	H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	459	2	Hamburger	Hamburger	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	459	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	459	4	S.	S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	459	5	Crain	Crain	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	459	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	460	1	1984	1984	CD
work_lowlmiflivaeblgbyxkxrbt3nq	460	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	461	1	Acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	461	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	461	3	cogni-	cogni-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	461	4	tive	tive	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	461	5	compiling	compiling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	461	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	462	1	Cognition	cognition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	462	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	462	3	17(2):85–136	17(2):85–136	CD
work_lowlmiflivaeblgbyxkxrbt3nq	462	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	463	1	R.P.	R.P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	464	1	Higginson	Higginson	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	464	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	465	1	1987	1987	CD
work_lowlmiflivaeblgbyxkxrbt3nq	465	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	466	1	Fixing	fix	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	466	2	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	466	3	Assimilation	assimilation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	466	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	466	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	466	6	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	466	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	467	1	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	467	2	Microfilms	Microfilms	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	467	3	International	International	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	467	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	468	1	M.H.	M.H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	469	1	Jones	Jones	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	469	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	469	3	E.C.	E.C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	470	1	Carterette	Carterette	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	470	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	471	1	1963	1963	CD
work_lowlmiflivaeblgbyxkxrbt3nq	471	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	472	1	Redundancy	redundancy	NN
work_lowlmiflivaeblgbyxkxrbt3nq	472	2	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	472	3	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	472	4	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	472	5	free	free	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	472	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	472	7	reading	read	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	472	8	choices	choice	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	472	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	473	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	473	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	473	3	Verbal	Verbal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	473	4	Learning	Learning	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	473	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	473	6	Verbal	Verbal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	473	7	Behavior	Behavior	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	473	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	473	9	2(5	2(5	NN
work_lowlmiflivaeblgbyxkxrbt3nq	473	10	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	473	11	6):489–493	6):489–493	CD
work_lowlmiflivaeblgbyxkxrbt3nq	473	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	474	1	P.	P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	2	Koehn	Koehn	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	4	H.	H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	5	Hoang	Hoang	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	7	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	8	Birch	Birch	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	10	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	11	Callison	Callison	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	12	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	474	13	Burch	Burch	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	15	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	16	Federico	Federico	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	18	N.	N.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	19	Bertoldi	Bertoldi	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	20	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	21	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	22	Cowan	Cowan	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	23	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	24	W.	W.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	25	Shen	Shen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	26	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	27	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	28	Moran	Moran	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	29	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	30	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	31	Zens	Zens	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	32	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	474	33	et	et	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	34	al	al	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	474	35	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	475	1	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	475	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	476	1	Moses	Moses	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	476	2	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	476	3	Open	open	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	476	4	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	476	5	toolkit	toolkit	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	476	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	476	7	statistical	statistical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	476	8	machine	machine	NN
work_lowlmiflivaeblgbyxkxrbt3nq	476	9	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	476	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	477	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	477	2	Proceed-	Proceed-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	3	ings	ing	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	477	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	477	5	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	477	6	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	7	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	477	8	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	9	Linguis-	Linguis-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	10	tics	tic	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	477	11	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	477	12	Interactive	Interactive	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	13	Poster	Poster	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	14	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	477	15	Demonstration	Demonstration	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	477	16	Sessions	Sessions	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	477	17	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	477	18	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	477	19	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	477	20	177–180	177–180	CD
work_lowlmiflivaeblgbyxkxrbt3nq	477	21	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	478	1	S.	S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	478	2	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	478	3	Kuczaj	Kuczaj	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	478	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	479	1	1977	1977	CD
work_lowlmiflivaeblgbyxkxrbt3nq	479	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	480	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	480	2	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	480	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	480	4	regular	regular	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	480	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	480	6	irreg-	irreg-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	480	7	ular	ular	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	480	8	past	past	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	480	9	tense	tense	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	480	10	forms	form	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	480	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	481	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	481	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	481	3	Verbal	Verbal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	481	4	Learning	Learning	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	481	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	481	6	Verbal	Verbal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	481	7	Behavior	Behavior	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	481	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	481	9	16(5):589–600	16(5):589–600	CD
work_lowlmiflivaeblgbyxkxrbt3nq	481	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	482	1	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	482	2	Lee	Lee	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	482	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	482	4	S.	S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	482	5	Seneff	Seneff	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	482	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	483	1	2006	2006	CD
work_lowlmiflivaeblgbyxkxrbt3nq	483	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	484	1	Automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	484	2	grammar	grammar	NN
work_lowlmiflivaeblgbyxkxrbt3nq	484	3	cor-	cor-	RB
work_lowlmiflivaeblgbyxkxrbt3nq	484	4	rection	rection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	484	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	484	6	second	second	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	484	7	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	484	8	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	484	9	learners	learner	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	484	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	485	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	485	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	485	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	485	5	International	International	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	6	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	7	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	485	8	Spoken	Spoken	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	9	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	10	Processing	Processing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	485	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	486	1	X.	X.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	486	2	Lu	Lu	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	486	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	487	1	2009	2009	CD
work_lowlmiflivaeblgbyxkxrbt3nq	487	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	488	1	Automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	488	2	measurement	measurement	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	488	4	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	488	5	com-	com-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	6	plexity	plexity	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	488	8	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	9	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	10	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	488	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	489	1	International	International	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	489	2	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	489	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	489	4	Corpus	Corpus	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	489	5	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	489	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	489	7	14(1):3–28	14(1):3–28	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	489	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	490	1	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	490	2	MacWhinney	MacWhinney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	490	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	491	1	2000	2000	CD
work_lowlmiflivaeblgbyxkxrbt3nq	491	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	492	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	492	2	CHILDES	CHILDES	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	492	3	project	project	NN
work_lowlmiflivaeblgbyxkxrbt3nq	492	4	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	492	5	Tools	tool	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	492	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	492	7	analyzing	analyze	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	492	8	talk	talk	NN
work_lowlmiflivaeblgbyxkxrbt3nq	492	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	492	10	volume	volume	NN
work_lowlmiflivaeblgbyxkxrbt3nq	492	11	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	492	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	493	1	Psychology	Psychology	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	493	2	Press	Press	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	493	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	494	1	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	494	2	MacWhinney	MacWhinney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	494	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	495	1	2007	2007	CD
work_lowlmiflivaeblgbyxkxrbt3nq	495	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	496	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	496	2	TalkBank	TalkBank	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	496	3	project	project	NN
work_lowlmiflivaeblgbyxkxrbt3nq	496	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	497	1	Cre-	cre-	XX
work_lowlmiflivaeblgbyxkxrbt3nq	497	2	ating	ate	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	497	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	497	4	digitizing	digitize	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	497	5	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	497	6	corpora	corpora	NN
work_lowlmiflivaeblgbyxkxrbt3nq	497	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	497	8	Synchronic	Synchronic	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	497	9	Databases	Databases	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	497	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	497	11	1:163–180	1:163–180	CD
work_lowlmiflivaeblgbyxkxrbt3nq	497	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	498	1	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	498	2	Mohri	Mohri	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	498	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	499	1	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	499	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	500	1	System	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	500	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	500	3	method	method	NN
work_lowlmiflivaeblgbyxkxrbt3nq	500	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	500	5	epsilon	epsilon	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	500	6	removal	removal	NN
work_lowlmiflivaeblgbyxkxrbt3nq	500	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	500	8	weighted	weighted	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	500	9	automata	automata	NN
work_lowlmiflivaeblgbyxkxrbt3nq	500	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	500	11	transducers	transducer	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	500	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	500	13	June	June	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	500	14	3	3	CD
work_lowlmiflivaeblgbyxkxrbt3nq	500	15	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	501	1	US	US	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	501	2	Patent	Patent	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	501	3	7,383,185	7,383,185	CD
work_lowlmiflivaeblgbyxkxrbt3nq	501	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	502	1	E.	E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	502	2	Morley	Morley	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	502	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	502	4	E.	E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	502	5	Prud’hommeaux	Prud’hommeaux	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	502	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	503	1	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	503	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	504	1	Using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	504	2	con-	con-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	504	3	stituency	stituency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	504	4	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	504	5	dependency	dependency	NN
work_lowlmiflivaeblgbyxkxrbt3nq	504	6	parse	parse	NN
work_lowlmiflivaeblgbyxkxrbt3nq	504	7	features	feature	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	504	8	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	504	9	identify	identify	VB
work_lowlmiflivaeblgbyxkxrbt3nq	504	10	er-	er-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	504	11	rorful	rorful	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	504	12	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	504	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	504	14	disordered	disorder	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	504	15	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	504	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	505	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	505	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	505	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	505	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	505	5	Workshop	Workshop	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	505	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	505	7	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	505	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	505	9	Computer	Computer	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	505	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	505	11	Interaction	Interaction	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	505	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	506	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	506	2	Ninio	Ninio	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	506	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	506	4	C.E.	C.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	507	1	Snow	Snow	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	507	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	507	3	B.A.	B.A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	508	1	Pan	Pan	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	508	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	508	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	508	4	P.R.	P.R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	509	1	Rollins	rollin	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	509	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	510	1	1994	1994	CD
work_lowlmiflivaeblgbyxkxrbt3nq	510	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	511	1	Classifying	classify	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	511	2	communicative	communicative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	511	3	acts	act	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	511	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	511	5	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	511	6	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	511	7	interactions	interaction	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	511	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	512	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	512	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	512	3	Communication	Communication	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	512	4	Disorders	Disorders	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	512	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	512	6	27(2):157–187	27(2):157–187	CD
work_lowlmiflivaeblgbyxkxrbt3nq	512	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	513	1	F.J.	F.J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	513	2	Och	Och	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	513	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	513	4	H.	H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	513	5	Ney	Ney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	513	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	514	1	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	514	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	515	1	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	515	2	systematic	systematic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	515	3	comparison	comparison	NN
work_lowlmiflivaeblgbyxkxrbt3nq	515	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	515	5	various	various	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	515	6	statistical	statistical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	515	7	alignment	alignment	NN
work_lowlmiflivaeblgbyxkxrbt3nq	515	8	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	515	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	516	1	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	516	2	Linguistics	Linguistics	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	516	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	516	4	29(1):19–51	29(1):19–51	CD
work_lowlmiflivaeblgbyxkxrbt3nq	516	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	517	1	F.J.	F.J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	517	2	Och	Och	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	517	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	518	1	2003	2003	CD
work_lowlmiflivaeblgbyxkxrbt3nq	518	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	519	1	Minimum	minimum	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	519	2	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	519	3	rate	rate	NN
work_lowlmiflivaeblgbyxkxrbt3nq	519	4	training	training	NN
work_lowlmiflivaeblgbyxkxrbt3nq	519	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	519	6	statistical	statistical	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	519	7	machine	machine	NN
work_lowlmiflivaeblgbyxkxrbt3nq	519	8	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	519	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	520	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	520	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	520	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	520	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	520	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	520	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	520	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	520	8	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	520	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	520	10	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	520	11	160–167	160–167	CD
work_lowlmiflivaeblgbyxkxrbt3nq	520	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	521	1	R.E.	R.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	522	1	Owens	owen	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	522	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	523	1	2008	2008	CD
work_lowlmiflivaeblgbyxkxrbt3nq	523	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	524	1	Language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	524	2	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	524	3	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	524	4	An	an	DT
work_lowlmiflivaeblgbyxkxrbt3nq	524	5	intro-	intro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	524	6	duction	duction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	524	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	525	1	Pearson	Pearson	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	2	Education	Education	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	525	4	Inc.	Inc.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	5	136	136	CD
work_lowlmiflivaeblgbyxkxrbt3nq	525	6	K.	K.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	7	Papineni	Papineni	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	525	9	S.	S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	10	Roukos	Roukos	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	525	12	T.	T.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	13	Ward	Ward	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	525	14	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	525	15	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	525	16	W.J.	W.J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	526	1	Zhu	zhu	FW
work_lowlmiflivaeblgbyxkxrbt3nq	526	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	527	1	2002	2002	CD
work_lowlmiflivaeblgbyxkxrbt3nq	527	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	528	1	BLEU	BLEU	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	528	2	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	528	3	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	528	4	method	method	NN
work_lowlmiflivaeblgbyxkxrbt3nq	528	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	528	6	automatic	automatic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	528	7	evaluation	evaluation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	528	8	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	528	9	machine	machine	NN
work_lowlmiflivaeblgbyxkxrbt3nq	528	10	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	528	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	529	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	529	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	529	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	529	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	529	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	529	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	529	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	529	8	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	529	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	529	10	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	529	11	311–318	311–318	CD
work_lowlmiflivaeblgbyxkxrbt3nq	529	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	530	1	Y.A.	Y.A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	531	1	Park	Park	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	531	2	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	531	3	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	531	4	Levy	Levy	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	531	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	532	1	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	532	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	533	1	Automated	automate	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	533	2	whole	whole	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	533	3	sentence	sentence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	533	4	grammar	grammar	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	533	5	correction	correction	NN
work_lowlmiflivaeblgbyxkxrbt3nq	533	6	using	use	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	533	7	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	533	8	noisy	noisy	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	533	9	channel	channel	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	533	10	model	model	NN
work_lowlmiflivaeblgbyxkxrbt3nq	533	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	534	1	Pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	534	2	ceedings	ceeding	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	534	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	534	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	534	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	534	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	534	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	534	8	Lin-	Lin-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	534	9	guistics	guistic	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	534	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	534	11	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	534	12	934–944	934–944	CD
work_lowlmiflivaeblgbyxkxrbt3nq	534	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	535	1	A.M.	A.M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	535	2	Peters	Peters	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	535	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	536	1	1987	1987	CD
work_lowlmiflivaeblgbyxkxrbt3nq	536	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	537	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	537	2	role	role	NN
work_lowlmiflivaeblgbyxkxrbt3nq	537	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	537	4	imitation	imitation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	537	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	537	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	537	7	devel-	devel-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	537	8	oping	oping	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	537	9	syntax	syntax	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	537	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	537	11	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	537	12	blind	blind	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	537	13	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	537	14	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	537	15	perspectives	perspective	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	537	16	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	537	17	repeti-	repeti-	FW
work_lowlmiflivaeblgbyxkxrbt3nq	537	18	tion	tion	NN
work_lowlmiflivaeblgbyxkxrbt3nq	537	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	538	1	Text	text	NN
work_lowlmiflivaeblgbyxkxrbt3nq	538	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	538	3	7(3):289–311	7(3):289–311	CD
work_lowlmiflivaeblgbyxkxrbt3nq	538	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	539	1	K.	K.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	539	2	Post	Post	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	539	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	540	1	1992	1992	CD
work_lowlmiflivaeblgbyxkxrbt3nq	540	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	541	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	541	2	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	541	3	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	541	4	environment	environment	NN
work_lowlmiflivaeblgbyxkxrbt3nq	541	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	541	6	laterborns	laterborn	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	541	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	541	8	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	541	9	rural	rural	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	541	10	Florida	Florida	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	541	11	community	community	NN
work_lowlmiflivaeblgbyxkxrbt3nq	541	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	542	1	Ph.D.	Ph.D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	542	2	thesis	thesis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	542	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	542	4	Harvard	Harvard	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	542	5	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	542	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	543	1	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	2	Quirk	Quirk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	543	4	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	5	Brockett	Brockett	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	543	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	543	8	W.	W.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	9	Dolan	Dolan	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	543	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	544	1	2004	2004	CD
work_lowlmiflivaeblgbyxkxrbt3nq	544	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	545	1	Monolin-	Monolin-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	545	2	gual	gual	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	545	3	machine	machine	NN
work_lowlmiflivaeblgbyxkxrbt3nq	545	4	translation	translation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	545	5	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	545	6	paraphrase	paraphrase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	545	7	generation	generation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	545	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	546	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	546	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	546	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	546	5	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	546	7	Empirical	Empirical	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	8	Methods	Methods	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	546	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	546	10	Natural	Natural	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	11	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	12	Processing	Processing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	546	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	546	14	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	546	15	142–149	142–149	CD
work_lowlmiflivaeblgbyxkxrbt3nq	546	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	547	1	T.	T.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	547	2	Regier	Regier	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	547	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	548	1	2005	2005	CD
work_lowlmiflivaeblgbyxkxrbt3nq	548	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	549	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	549	2	emergence	emergence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	549	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	549	4	words	word	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	549	5	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	549	6	Attentional	attentional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	549	7	learning	learning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	549	8	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	549	9	form	form	NN
work_lowlmiflivaeblgbyxkxrbt3nq	549	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	549	11	meaning	meaning	NN
work_lowlmiflivaeblgbyxkxrbt3nq	549	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	550	1	Cognitive	Cognitive	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	550	2	Science	Science	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	550	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	550	4	29(6):819–865	29(6):819–865	CD
work_lowlmiflivaeblgbyxkxrbt3nq	550	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	551	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	551	2	Ritter	Ritter	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	551	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	551	4	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	551	5	Cherry	Cherry	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	551	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	551	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	551	8	W.B.	W.B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	552	1	Dolan	Dolan	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	552	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	553	1	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	553	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	554	1	Data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	554	2	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	554	3	driven	drive	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	554	4	response	response	NN
work_lowlmiflivaeblgbyxkxrbt3nq	554	5	generation	generation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	554	6	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	554	7	social	social	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	554	8	media	medium	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	554	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	555	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	555	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	555	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	555	5	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	555	7	Empirical	Empirical	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	8	Methods	Methods	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	555	9	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	555	10	Natural	Natural	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	11	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	12	Processing	Processing	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	555	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	555	14	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	555	15	583–593	583–593	CD
work_lowlmiflivaeblgbyxkxrbt3nq	555	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	556	1	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	2	Roark	Roark	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	556	4	R.	R.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	5	Sproat	Sproat	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	556	7	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	8	Allauzen	Allauzen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	556	10	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	11	Riley	Riley	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	12	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	556	13	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	14	Sorensen	Sorensen	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	15	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	556	16	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	556	17	T.	T.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	18	Tai	Tai	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	556	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	557	1	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	557	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	558	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	558	2	OpenGrm	OpenGrm	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	558	3	open	open	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	558	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	558	5	source	source	NN
work_lowlmiflivaeblgbyxkxrbt3nq	558	6	finite-	finite-	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	558	7	state	state	NN
work_lowlmiflivaeblgbyxkxrbt3nq	558	8	grammar	grammar	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	558	9	software	software	NN
work_lowlmiflivaeblgbyxkxrbt3nq	558	10	libraries	library	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	558	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	559	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	559	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	559	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	559	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	559	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	559	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	559	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	559	8	Linguistics	Linguistics	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	559	9	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	559	10	System	System	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	559	11	Demonstrations	Demonstrations	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	559	12	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	559	13	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	559	14	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	559	15	61–66	61–66	CD
work_lowlmiflivaeblgbyxkxrbt3nq	559	16	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	560	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	2	Rozovskaya	Rozovskaya	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	560	4	M.	M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	5	Sammons	Sammons	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	560	7	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	8	Gioja	Gioja	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	560	10	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	560	11	D.	D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	12	Roth	Roth	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	560	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	561	1	2011	2011	CD
work_lowlmiflivaeblgbyxkxrbt3nq	561	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	562	1	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	562	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	562	3	Illinois	Illinois	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	562	4	system	system	NN
work_lowlmiflivaeblgbyxkxrbt3nq	562	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	562	6	HOO	HOO	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	562	7	text	text	NN
work_lowlmiflivaeblgbyxkxrbt3nq	562	8	cor-	cor-	IN
work_lowlmiflivaeblgbyxkxrbt3nq	562	9	rection	rection	NN
work_lowlmiflivaeblgbyxkxrbt3nq	562	10	shared	share	VBD
work_lowlmiflivaeblgbyxkxrbt3nq	562	11	task	task	NN
work_lowlmiflivaeblgbyxkxrbt3nq	562	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	563	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	563	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	563	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	563	5	European	European	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	6	Workshop	Workshop	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	7	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	563	8	Natural	Natural	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	9	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	10	Generation	Generation	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	563	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	563	12	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	563	13	263–266	263–266	CD
work_lowlmiflivaeblgbyxkxrbt3nq	563	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	564	1	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	564	2	Sachs	Sachs	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	564	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	565	1	1983	1983	CD
work_lowlmiflivaeblgbyxkxrbt3nq	565	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	566	1	Talking	talk	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	566	2	about	about	IN
work_lowlmiflivaeblgbyxkxrbt3nq	566	3	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	566	4	there	there	RB
work_lowlmiflivaeblgbyxkxrbt3nq	566	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	566	6	then	then	RB
work_lowlmiflivaeblgbyxkxrbt3nq	566	7	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	566	8	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	566	9	emergence	emergence	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	10	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	566	11	displaced	displace	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	566	12	reference	reference	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	13	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	566	14	parent	parent	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	15	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	566	16	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	17	dis-	dis-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	18	course	course	NN
work_lowlmiflivaeblgbyxkxrbt3nq	566	19	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	567	1	Children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	567	2	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	567	3	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	567	4	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	567	5	4	4	CD
work_lowlmiflivaeblgbyxkxrbt3nq	567	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	568	1	K.	K.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	568	2	Sagae	Sagae	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	568	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	568	4	A.	a.	NN
work_lowlmiflivaeblgbyxkxrbt3nq	568	5	Lavie	Lavie	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	568	6	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	568	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	568	8	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	568	9	MacWhinney	MacWhinney	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	568	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	569	1	2005	2005	CD
work_lowlmiflivaeblgbyxkxrbt3nq	569	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	570	1	Auto-	Auto-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	570	2	matic	matic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	570	3	measurement	measurement	NN
work_lowlmiflivaeblgbyxkxrbt3nq	570	4	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	570	5	syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	570	6	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	570	7	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	570	8	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	570	9	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	570	10	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	571	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	571	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	571	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	571	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	571	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	571	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	571	7	Com-	Com-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	571	8	putational	putational	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	571	9	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	571	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	571	11	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	571	12	197–204	197–204	CD
work_lowlmiflivaeblgbyxkxrbt3nq	571	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	572	1	S.	S.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	572	2	Sahakian	Sahakian	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	572	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	572	4	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	572	5	Snyder	Snyder	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	572	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	573	1	2012	2012	CD
work_lowlmiflivaeblgbyxkxrbt3nq	573	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	574	1	Automatically	automatically	RB
work_lowlmiflivaeblgbyxkxrbt3nq	574	2	learn-	learn-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	574	3	ing	ing	NN
work_lowlmiflivaeblgbyxkxrbt3nq	574	4	measures	measure	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	574	5	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	574	6	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	574	7	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	574	8	development	development	NN
work_lowlmiflivaeblgbyxkxrbt3nq	574	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	575	1	Pro-	pro-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	575	2	ceedings	ceeding	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	575	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	575	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	575	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	575	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	575	7	Computational	Computational	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	575	8	Lin-	Lin-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	575	9	guistics	guistic	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	575	10	(	(	-LRB-
work_lowlmiflivaeblgbyxkxrbt3nq	575	11	Volume	volume	NN
work_lowlmiflivaeblgbyxkxrbt3nq	575	12	2	2	CD
work_lowlmiflivaeblgbyxkxrbt3nq	575	13	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	575	14	Short	Short	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	575	15	Papers	Papers	NNPS
work_lowlmiflivaeblgbyxkxrbt3nq	575	16	)	)	-RRB-
work_lowlmiflivaeblgbyxkxrbt3nq	575	17	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	575	18	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	575	19	95–99	95–99	CD
work_lowlmiflivaeblgbyxkxrbt3nq	575	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	576	1	C.E.	C.E.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	1	Snow	Snow	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	577	3	F.	F.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	4	Shonkoff	Shonkoff	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	577	6	K.	K.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	7	Lee	Lee	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	577	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	577	10	H.	H.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	11	Levin	Levin	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	577	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	578	1	1986	1986	CD
work_lowlmiflivaeblgbyxkxrbt3nq	578	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	579	1	Learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	579	2	to	to	TO
work_lowlmiflivaeblgbyxkxrbt3nq	579	3	play	play	VB
work_lowlmiflivaeblgbyxkxrbt3nq	579	4	doctor	doctor	NN
work_lowlmiflivaeblgbyxkxrbt3nq	579	5	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	579	6	Effects	effect	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	579	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	579	8	sex	sex	NN
work_lowlmiflivaeblgbyxkxrbt3nq	579	9	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	579	10	age	age	NN
work_lowlmiflivaeblgbyxkxrbt3nq	579	11	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	579	12	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	579	13	ex-	ex-	DT
work_lowlmiflivaeblgbyxkxrbt3nq	579	14	perience	perience	NN
work_lowlmiflivaeblgbyxkxrbt3nq	579	15	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	579	16	hospital	hospital	NN
work_lowlmiflivaeblgbyxkxrbt3nq	579	17	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	580	1	Discourse	Discourse	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	580	2	Processes	Processes	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	580	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	580	4	9(4):461	9(4):461	CD
work_lowlmiflivaeblgbyxkxrbt3nq	580	5	–	–	:
work_lowlmiflivaeblgbyxkxrbt3nq	580	6	473	473	CD
work_lowlmiflivaeblgbyxkxrbt3nq	580	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	581	1	E.L.	E.L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	581	2	Stine	Stine	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	581	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	581	4	J.N.	J.N.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	582	1	Bohannon	Bohannon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	582	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	583	1	1983	1983	CD
work_lowlmiflivaeblgbyxkxrbt3nq	583	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	584	1	Imitations	imitation	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	584	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	584	3	inter-	inter-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	584	4	actions	action	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	584	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	584	6	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	584	7	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	584	8	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	584	9	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	585	1	Journal	Journal	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	585	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	585	3	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	585	4	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	585	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	585	6	10(03):589–603	10(03):589–603	CD
work_lowlmiflivaeblgbyxkxrbt3nq	585	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	586	1	X.	X.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	1	Sun	Sun	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	587	3	J.	J.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	4	Gao	Gao	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	5	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	587	6	D.	D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	7	Micol	Micol	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	8	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	587	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	587	10	C.	C.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	11	Quirk	Quirk	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	587	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	588	1	2010	2010	CD
work_lowlmiflivaeblgbyxkxrbt3nq	588	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	589	1	Learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	589	2	phrase	phrase	NN
work_lowlmiflivaeblgbyxkxrbt3nq	589	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	589	4	based	base	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	589	5	spelling	spelling	NN
work_lowlmiflivaeblgbyxkxrbt3nq	589	6	error	error	NN
work_lowlmiflivaeblgbyxkxrbt3nq	589	7	models	model	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	589	8	from	from	IN
work_lowlmiflivaeblgbyxkxrbt3nq	589	9	clickthrough	clickthrough	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	589	10	data	datum	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	589	11	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	590	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	590	2	Proceedings	Proceedings	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	590	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	590	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	590	5	Association	Association	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	590	6	for	for	IN
work_lowlmiflivaeblgbyxkxrbt3nq	590	7	Computa-	Computa-	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	590	8	tional	tional	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	590	9	Linguistics	Linguistics	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	590	10	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	590	11	pages	page	VBZ
work_lowlmiflivaeblgbyxkxrbt3nq	590	12	266–274	266–274	CD
work_lowlmiflivaeblgbyxkxrbt3nq	590	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	591	1	P.	P.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	591	2	Suppes	Suppes	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	591	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	592	1	1974	1974	CD
work_lowlmiflivaeblgbyxkxrbt3nq	592	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	593	1	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	593	2	semantics	semantic	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	593	3	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	593	4	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	593	5	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	593	6	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	593	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	594	1	American	American	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	594	2	Psychologist	Psychologist	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	594	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	594	4	29(2):103	29(2):103	CD
work_lowlmiflivaeblgbyxkxrbt3nq	594	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	595	1	T.Z.	T.Z.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	596	1	Tardif	Tardif	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	596	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	597	1	1994	1994	CD
work_lowlmiflivaeblgbyxkxrbt3nq	597	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	598	1	Adult	adult	NN
work_lowlmiflivaeblgbyxkxrbt3nq	598	2	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	598	3	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	598	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	598	5	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	598	6	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	598	7	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	598	8	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	598	9	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	598	10	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	598	11	Mandarin	Mandarin	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	598	12	Chinese	Chinese	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	598	13	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	599	1	Ph.D.	Ph.D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	599	2	thesis	thesis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	599	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	599	4	Yale	Yale	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	599	5	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	599	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	600	1	V.	V.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	600	2	Valian	Valian	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	600	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	601	1	1991	1991	CD
work_lowlmiflivaeblgbyxkxrbt3nq	601	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	602	1	Syntactic	syntactic	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	602	2	subjects	subject	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	602	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	602	4	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	602	5	early	early	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	602	6	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	602	7	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	602	8	American	american	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	602	9	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	602	10	Italian	italian	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	602	11	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	602	12	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	603	1	Cognition	cognition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	603	2	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	603	3	40(1	40(1	CD
work_lowlmiflivaeblgbyxkxrbt3nq	603	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	603	5	2):21	2):21	CD
work_lowlmiflivaeblgbyxkxrbt3nq	603	6	–	–	:
work_lowlmiflivaeblgbyxkxrbt3nq	603	7	81	81	CD
work_lowlmiflivaeblgbyxkxrbt3nq	603	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	604	1	L.	L.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	604	2	Van	Van	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	604	3	Houten	Houten	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	604	4	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	605	1	1986	1986	CD
work_lowlmiflivaeblgbyxkxrbt3nq	605	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	606	1	Role	role	NN
work_lowlmiflivaeblgbyxkxrbt3nq	606	2	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	606	3	maternal	maternal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	606	4	input	input	NN
work_lowlmiflivaeblgbyxkxrbt3nq	606	5	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	606	6	the	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	606	7	acquisition	acquisition	NN
work_lowlmiflivaeblgbyxkxrbt3nq	606	8	process	process	NN
work_lowlmiflivaeblgbyxkxrbt3nq	606	9	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	606	10	The	the	DT
work_lowlmiflivaeblgbyxkxrbt3nq	606	11	communicative	communicative	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	606	12	strategies	strategy	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	606	13	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	606	14	adolescent	adolescent	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	606	15	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	606	16	older	old	JJR
work_lowlmiflivaeblgbyxkxrbt3nq	606	17	mothers	mother	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	606	18	with	with	IN
work_lowlmiflivaeblgbyxkxrbt3nq	606	19	their	-PRON-	PRP$
work_lowlmiflivaeblgbyxkxrbt3nq	606	20	language	language	NN
work_lowlmiflivaeblgbyxkxrbt3nq	606	21	learning	learn	VBG
work_lowlmiflivaeblgbyxkxrbt3nq	606	22	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	606	23	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	607	1	In	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	607	2	Boston	Boston	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	607	3	University	University	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	607	4	Conference	Conference	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	607	5	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	607	6	Language	Language	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	607	7	Development	Development	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	607	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	608	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	608	2	Warren	Warren	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	608	3	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	608	4	Leubecker	Leubecker	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	608	5	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	608	6	J.N.	J.N.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	609	1	Bohannon	Bohannon	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	609	2	III	III	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	609	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	610	1	1984	1984	CD
work_lowlmiflivaeblgbyxkxrbt3nq	610	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	611	1	Into-	into-	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	611	2	nation	nation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	611	3	patterns	pattern	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	611	4	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	611	5	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	611	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	611	7	directed	direct	VBN
work_lowlmiflivaeblgbyxkxrbt3nq	611	8	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	611	9	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	611	10	Mother	mother	NN
work_lowlmiflivaeblgbyxkxrbt3nq	611	11	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	611	12	father	father	NN
work_lowlmiflivaeblgbyxkxrbt3nq	611	13	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	611	14	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	612	1	Child	Child	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	612	2	Development	Development	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	612	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	612	4	55(4):1379–1385	55(4):1379–1385	CD
work_lowlmiflivaeblgbyxkxrbt3nq	612	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	613	1	A.	A.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	613	2	Warren	Warren	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	613	3	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	614	1	1982	1982	CD
work_lowlmiflivaeblgbyxkxrbt3nq	614	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	615	1	Sex	sex	NN
work_lowlmiflivaeblgbyxkxrbt3nq	615	2	differences	difference	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	615	3	in	in	IN
work_lowlmiflivaeblgbyxkxrbt3nq	615	4	speech	speech	NN
work_lowlmiflivaeblgbyxkxrbt3nq	615	5	to	to	IN
work_lowlmiflivaeblgbyxkxrbt3nq	615	6	children	child	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	615	7	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	616	1	Ph.D.	Ph.D.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	616	2	thesis	thesis	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	616	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	616	4	Georgia	Georgia	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	616	5	Institute	Institute	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	616	6	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	616	7	Technology	Technology	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	616	8	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	617	1	B.	B.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	617	2	Wilson	Wilson	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	617	3	and	and	CC
work_lowlmiflivaeblgbyxkxrbt3nq	617	4	A.M.	A.M.	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	617	5	Peters	Peters	NNP
work_lowlmiflivaeblgbyxkxrbt3nq	617	6	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	618	1	1988	1988	CD
work_lowlmiflivaeblgbyxkxrbt3nq	618	2	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	619	1	What	what	WP
work_lowlmiflivaeblgbyxkxrbt3nq	619	2	are	be	VBP
work_lowlmiflivaeblgbyxkxrbt3nq	619	3	you	-PRON-	PRP
work_lowlmiflivaeblgbyxkxrbt3nq	619	4	cookin	cookin	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	619	5	’	'	''
work_lowlmiflivaeblgbyxkxrbt3nq	619	6	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	619	7	a	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	619	8	hot	hot	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	619	9	?	?	.
work_lowlmiflivaeblgbyxkxrbt3nq	620	1	:	:	:
work_lowlmiflivaeblgbyxkxrbt3nq	620	2	A	a	DT
work_lowlmiflivaeblgbyxkxrbt3nq	620	3	three	three	CD
work_lowlmiflivaeblgbyxkxrbt3nq	620	4	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	620	5	year	year	NN
work_lowlmiflivaeblgbyxkxrbt3nq	620	6	-	-	HYPH
work_lowlmiflivaeblgbyxkxrbt3nq	620	7	old	old	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	620	8	blind	blind	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	620	9	child	child	NN
work_lowlmiflivaeblgbyxkxrbt3nq	620	10	’s	’s	POS
work_lowlmiflivaeblgbyxkxrbt3nq	620	11	‘	'	``
work_lowlmiflivaeblgbyxkxrbt3nq	620	12	violation	violation	NN
work_lowlmiflivaeblgbyxkxrbt3nq	620	13	’	'	''
work_lowlmiflivaeblgbyxkxrbt3nq	620	14	of	of	IN
work_lowlmiflivaeblgbyxkxrbt3nq	620	15	universal	universal	JJ
work_lowlmiflivaeblgbyxkxrbt3nq	620	16	constraints	constraint	NNS
work_lowlmiflivaeblgbyxkxrbt3nq	620	17	on	on	IN
work_lowlmiflivaeblgbyxkxrbt3nq	620	18	constituent	constituent	NN
work_lowlmiflivaeblgbyxkxrbt3nq	620	19	movement	movement	NN
work_lowlmiflivaeblgbyxkxrbt3nq	620	20	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	621	1	Lan-	lan-	NN
work_lowlmiflivaeblgbyxkxrbt3nq	621	2	guage	guage	NN
work_lowlmiflivaeblgbyxkxrbt3nq	621	3	,	,	,
work_lowlmiflivaeblgbyxkxrbt3nq	621	4	64:249–273	64:249–273	CD
work_lowlmiflivaeblgbyxkxrbt3nq	621	5	.	.	.
work_lowlmiflivaeblgbyxkxrbt3nq	622	1	137	137	CD
work_lowlmiflivaeblgbyxkxrbt3nq	622	2	138	138	CD