Other so-called invariant question tags such as right,yes, no and eh fulfil a similar function.
In section two, I report three pieces of evidence discussed in Palacios Martinez (2011, 2012 and 2013) that show that innit departs from the usual behaviour of other question tags in English.
This is shown in (7), where some of the sentences with innit from (4) have been paraphrased with canonical question tags.
2011; among others), do not exist as a separate class of question tags.
For all the reasons discussed above, and given the striking differences between the behaviour of canonical question tags and that of innit, I conclude that it is more desirable to put forward an analysis of innit within what we know about the syntax of pragmatic particles than doing so within the syntax of question tags.
I also show that an analysis along the lines of Krifka's makes it possible to explain why declarative sentences with innit can be interpreted not only as declaratives with canonical question tags, but also as negated polarity questions with high negation.
Krifka (forthcoming, 27) argues that questions containing question tags, which are biased questions, involve the speaker's commitment to the truth of the proposition and restrict the addressee's reaction to either accepting p as part of the CG, or to asserting notp without having to reject treatingp as CG.
I will assume, in line with Sailor (2011) and Barros and Craenenbroeck (2013), that questions with question tags involve vp ellipsis.
4) In addition to the ten features noted by Janda, the following were found to be salient in this study: deleted "that" complementizers, reflexives, causals, discourse markers, question tags, process statements, contracted negatives, and negative statements transformed into positives.
For example, the main idea components of talk, such as topics and responses, are more central to note preservation than are small talk, formulaic greetings and closings, timing indications, meta-language such as process statements, question tags, discourse markers, and address forms.
Other discourse violations, including timing indicators, process statements, discourse markers, requests for clarification, question tags, and address forms, do not occur as frequently and, by themselves, may not appear to be strong e vidence of an underlying tape recording.