I am currently working on fixing a bug in BritishReplaceRule (and other identical rules). The problem is that 2-word American expressions are not detected by LanguageTool. I studied the code and I found that the cause of the problem is detected in the lines below (from the class AbstractSimpleReplaceRule):
As each tokenReading in the tokens array refers to a single word of the sentence.
So my question is: is there a way that two AnalyzedTokenReadings can be connected (in order to refer to 2 words)?