Antidote 10 User Guide

User Guide / The Corrector / Understanding How the Corrector Works

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.

Discontinuities

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.

Unknown words

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.