A/an/the articles in English suggestion/correction

1

Hi,

Is it possible to analyze text with langtool such that missing or unnecessary articles are alerted ?
I am using command line version with Emacs.

Thanks,

2

There are several rules in LT that try to detect issues with articles. You can see them at Browse LanguageTool Rules: 203 matches for English. If you only want to find these errors, you’d have to turn on only these rules by their id (parameters --enabledonly --enable ID1,ID2,... when using the LT command line version).

3

I am not an expert, but I have two examples

You eat an apple. which is correct
LanguageTool will correct me when I write
You eat a apple
but not
You eat apple. this one seems not correct

Could LanguageTool catch also missing articles from the context ?

I think there are three errors that one can commit

  1. Wrong article (Handled by LanguageTool)
  2. Missing article (Not detected)
  3. Unnecessary article (Not detected)

Suggestions that article is missing, would be nice.

4

Added to my task list. Thanks for the feedback.

5

Hi Mike,

That is really great, just curiosity… how long is the task list and how far are the articles ? :slight_smile:

Regards,

6

@OneSploit, to see the tasks that are assigned to a person on the Issues list (https://github.com/languagetool-org/languagetool/issues?utf8=%E2%9C%93&q=), filter the Assignee.

But, there are other tasks (like this one) that are not on that list.

I’m overloaded with stuff to do, so I probably will not do the rules for the articles until October. Sorry.

7

Hi Mike,

It’s ok for me no rush, I was just curious as it would help me a lot in my writing now but I would love to use it in the future as well. I am really grateful you are willing to look into this.

Thanks again,
Mateusz

8

Hi @OneSploit,

I added a rule ARTICLE_UNNECESSARY (https://github.com/languagetool-org/languagetool/commit/f49647f9058720052bcb5d402cf48bdd632eeb8e. The rule gives too many false positives because many nouns that have only the postag NN:U can also be NN, but they do not have the postag NN. Thus, they cannot be excluded using an exception. The rule is off by default.

For rules for missing articles, Refer to the rulegroup ARTICLE_MISSING in grammar.xml. The rulegroup has default=“off”. To use the rule, you must enable the rule.

When the rulegroup is enabled, the first rule gives an error message for the second sentence:
Correct: I need oil for my car.
Not correct: I need radiator for my car.

LT does not give a message for ‘You eat apple’, because ‘apple’ is both a count noun and an uncountable noun.