Understanding How the Corrector Works
The corrector carries out an in-depth analysis of every sentence in your text. It first identifies each word along with its syntactic category, number, tense, etc. It then determines all the relevant grammatical links between the words: subject–verb, verb–object, etc. Once that is done, Antidote can then verify the appropriate agreement relationships between the words and propose any necessary corrections.
The following are some of the features of your text that Antidote automatically verifies: the correctness of various grammatical agreements (subject–verb, including agreement with coordinated subjects; articles and determiners with count and mass nouns), verbal mood, the written form of numbers, the use of hyphens, false friends, homophones, register, regionalisms, repetitions, capitalization, punctuation, etc.
- It should be noted that Antidote’s grammar checking is based on the syntactic, grammatical and orthographic properties of your text, not on any deep semantic factors. Antidote does not really understand the deep meaning of a sentence and can’t necessarily turn a poorly written text into a literary masterpiece.
Despite all of the tools at its disposal, the corrector may not be able to analyze every sentence it receives. It uses a dotted line to indicate the area where its analysis ends. Any detections that fall within this area are less reliable, and their tooltips display the warning Review, inviting you to pay special attention.
Antidote knows more words than most of us—its lexicon includes over 94,000 words—but English is a living entity! Like us, the corrector can analyze a sentence even if it contains a word it does not know. However, it is rarely able to propose a correction for that word. Share your knowledge with Antidote: add the unknown word to one of your personal dictionaries. When Antidote later corrects you on that word, you’ll be glad you did.
- To make things easier, all unknown words are grouped together in the Language view’s list of detections. They can also be displayed using the Rare words filter in the revision prism, under Vocabulary.
The impact of the corrector on the Style view and the other prisms
Antidote uses the results of the corrector’s analysis to determine as precisely as possible all the detections in the Style view and the results in the various revision, statistical and inspection filters. This analysis is what allows it to distinguish, for example, a number that is part of a date from a number denoting a quantity. It is also why it can process all the sentences in your text at once and detect whole passages, not just individual words, and even flag certain pragmatic elements.
The quality of the detections highlighted in the Style view therefore depends directly on the quality of the corrector’s analysis: if the analysis is weak, so too will be the detections. For example, an unknown proper noun will sometimes be classified as a place rather than a person, or a passive construction will not be properly interpreted.
For this reason, we strongly advise you to first tackle the corrections suggested in the Language view. Help Antidote resolve its analysis issues (for example, by adding unknown proper nouns to your personal dictionary), before proceeding to the verification of style and to the other prisms. The results will benefit from your efforts.