I first came across the topic of non-core arguments in 2010 when I attended a talk on personal datives in Southern American English at the Arizona Linguistics Circle 4 in Tucson. A personal dative is an optional pronoun that is coreferential with the subject in a sentence: for example, me in I wanna hear me a sad song (from Toby Keith's country song ‘Get My Drink On’). I realized then, somewhat to my own surprise, that personal datives also exist in my native variety of Lebanese Arabic. Given my background in syntax, I immediately began to wonder: how are personal datives licensed in a position where reflexive pronouns are expected? I went on to analyze the syntax of personal datives, as well as other types of non-core arguments that are licensed in Lebanese Arabic, and before long I found myself delving into the pragmatic functions of these datives and the attitudinal contributions they make to utterances. That was when I started to refer to them as ‘attitude datives’.

For my syntactic and pragmatic analyses, I initially relied on constructed examples, elicited examples, and attested data that I collected during fieldwork. Four points became clear in the process. First, attitude datives are not unique to Lebanese Arabic, and are in fact a regional feature of Levantine Arabic (that is, Jordanian, Lebanese, Palestinian, and Syrian Arabic) in general. They are also licensed in most, if not all, other Arabic dialects; however, cross-dialectal variations become more apparent once one moves outside the Levant.

Second, the pragmatic contributions of attitude datives depend crucially on who is saying what to whom, where, when, and to what end. This means that syntactic and semantic accounts of these datives may successfully answer questions about their distribution and overarching meanings, but may not capture the wide range of meanings and functions that they can bring to an interaction. Instead, we need an analysis that goes beyond the sentence level, examining the social functions of a sizeable corpus of attitude datives as they are employed in situated utterances. I quickly realized that elicited examples and a small number of attested data in limited settings just would not do the trick.