The CAD hypothesis holds that there is mapping between the three moral emotions (contempt, anger and disgust) and the three moral codes of community, autonomy and divinity. Different from previous designs to establish correlations between emotions and eliciting situations which instantiate moral codes, this paper takes a narratological approach to the CAD hypothesis by examining the relationships between the three moral emotions and moral judgment relating to the three moral codes in the context of eliciting situations. First, similarity data pertaining to eliciting situations were collected by using the Order k/n-1 with fixed K method. Second, the participants were instructed to write down both their responses and justifications of their responses to the eliciting situations. A narratological analysis of the justifications of responses show that they vary along three variables: narrator, character, and basis (mostly in the form of moral judgment). The descriptive statistics of participants’ responses and of their justifications show that more than a half of responses are in the categories of anger (24.8%), disgust (20.7), and contempt (7.7%) and that about 60% of justifications contain a component of moral judgment based on the three moral codes of autonomy (30.03%), divinity (18.1), and community (11.82%). Correspondence analyses among eliciting situations, emotional responses and the three variables of justifications, together with results from the Multidimensional Scaling analysis of the similarity data, show that the CAD hypothesis is largely supported if mappings are set between the emotions in question and moral judgment concerning the eliciting situations (the basis variable of justification) and that the hypothesis is conditioned by the variable of character.

This paper addresses how to identify and code lexical items which explicitly realize attitudinal meaning. Such items are the kind of linguistic resources about whose coding discourse analysts feel confident (Martin 2003), but these have unfortunately received less scholarly attention than indirect realizations of attitudinal meaning such as metaphors and factual statements. In the process of annotating a small-scale corpus of 15 short stories published in English, it is observed that words with connotation, which are not regarded attitudinal lexis, also express opinions; some lexical items make hybrid realizations of logically related attitudinal meanings; and some verbs, traditionally called speech act terms, construe attitudinal meaning in a different way. Drawing on data collected from the small-scale corpus, this paper firstly proposes a distinction between typical realization and combinational realization of attitudinal meaning to account for the differences between attitudinal lexis and words with connotation; and then puts forward two further distinctions between independent and correlational realization and between projected and projecting realization of attitudinal meaning. This helps to explain the differences between lexical items realizing one attitudinal meaning and words realizing two attitudinal meanings, and between verbs that indicate speech behavior and other lexical items which do not.

J. R. Martin is a leading scholar who has greatly developed the theoretical framework of systemic functional linguistics (SFL) over the past four decades. Some of these contributions, such as the systems of discourse semantics, the appraisal framework and genre relations have been widely applied in various areas of linguistic studies and language education. The educational linguistic model he and his colleagues have cultivated and developed has attracted the attention of more and more educators from different disciplines around the globe. In this interview, he firstly elaborates on the significance of the concepts of discourse semantics, grammatical metaphor and genre. Then he continues with applications of genre theory in secondary school education, language maintenance, the relation and collaboration between Legitimation Code Theory (LCT) and SFL, and how the two paradigms complement each other. Finally, he introduces some of his recent collaborations with grammarians of different languages.

This paper investigates the mathematic features of non-linear models and discusses the processing way of non-linear factors which contributes to the non-linearity of a nonlinear model. On the basis of the error definition, this paper puts forward a new adjustment criterion, SGPE. Last,this paper investigates the solution of a non-linear regression model in the non-linear model space and makes the comparison between the estimated values in non-linear model space and those in linear model space.