Back to LanguageTool Homepage - Privacy - Imprint

MANY_NN replacement contains a trailing whitespace

(Erik van Leeuwen) #1


I’ve executed an en-US grammar check for the following text (using the API):
“There are a few JavaScript XPath implementations.”

The response contains a match for the MANY_NN rule telling me to replace “few JavaScript” with "few " (notice the trailing whitespace). If I execute that replacement the sentence would contain a double whitespace between the words few and XPath.

The match response:

	"message": "Possible agreement error. The noun JavaScript seems to be countable; consider using: \"few \".",
	"shortMessage": "Grammatical problem",
	"replacements": [
			"value": "few "
	"offset": 12,
	"length": 14,
	"context": {
		"text": "There are a few JavaScript XPath implementations.",
		"offset": 12,
		"length": 14
	"sentence": "There are a few JavaScript XPath implementations.",
	"rule": {
		"id": "MANY_NN",
		"subId": "1",
		"description": "Possible agreement error: 'many/several/few' + singular countable noun",
		"issueType": "grammar",
		"category": {
			"id": "GRAMMAR",
			"name": "Grammar"

The MANY_NN rule message is:

<message>Possible agreement error. The noun <match no="2"/> seems to be countable; consider using: <suggestion><match no="1"/> <match suppress_misspelled="yes" no="2" postag="NNP?S" postag_regexp="yes"/></suggestion>.</message>

I think the whitespace in the <suggestion> between the <match no="1"/> and <match no="2"/> is causing this. Is this intended behavior? Is or should there be an option to only add the trailing whitespace when the <match no="2"/> is present?

(Daniel Naber) #2

The basic issue seems to be that a rule matches here that doesn’t consider the whole chunk of “JavaScript XPath implementations”. You can use to see how text is analyzed, and in this case, the chunk is detected correctly (I-NP-plural, I-NP-plural, E-NP-plural). So I think a better approach would be to fix the matching first before looking at the suggestions. You can look for chunk="..." in grammar.xml to see how chunks can be addressed in XML rules.