This book started out from the observation that the terms literal meaning and non-literal meaning – traditionally used in the characterisation of the difference between semantics and pragmatics – themselves are not clearly defined in the literature. On the contrary, it seems that they are implicitly assumed to be intuitively clear, basic notions that need no further characterisation. Nevertheless, from the way in which these and other terms were traditionally used in the characterisation of semantics and pragmatics, a kind of standard notion can be identified for each of them. In short these are such that literal meaning is characterised as context-independent, conventional and primary, whereas non-literal meaning is context-dependent, non-conventional and secondary to the former. In other words, the standard notions of these terms can be taken to derive from their use together with such terms as context- (in-)dependence and (non-)conventionality in the characterisation of semantics and pragmatics together with an assumption concerning the temporal order in which the processes within these two systems proceed.
However, as noted in chapter 1, examples for uses of the two terms can be found that do not conform to their standard characterisations. In fact, it is by authors who do not accept the traditional characterisation of the difference between semantics and pragmatics that the terms literal meaning and non-literal meaning are used in non-standard ways. The problem is that even in those approaches it is not explicitly made clear how exactly these latter two notions are to be understood. Be that as it may, the main point is that such non-standard uses of the terms literal meaning and non-literal meaning suggest that their standard characterisations may not adequately capture the characteristics of the kinds of meanings the terms are intuitively taken to refer to. In the present book, I followed this suggestion and the consequences it leads to. In particular, I aimed at answering the following questions.
- What is it that makes the standard notions of literal meaning and non-literal meaning inadequate and thus in need of revision?
- What exactly are the properties that characterise and differentiate literal meaning and non-literal meaning and how are these particular types of meaning related to other types of meaning identified in the semantics/pragmatics literature (e.g., conversational implicature, implicit meaning aspects)?
- By which criteria should semantics and pragmatics be characterised and differentiated, if not by the dichotomies traditionally used and under the assumption that the two systems are involved in the determination of (at least) three distinct meaning levels in interpretation?
I will summarise my findings below.
The standard notions of literal meaning and non-literal meaning were shown to be inadequate as they do not capture the kinds of meaning they intuitively are supposed to. In particular, the standard characterisation of literal meaning as essentially context-independent and easy to process and non-literal meaning as essentially non-conventional and hard to process is a misconception. Rather, it seems that even the literal meaning of an expression is not determined without a consideration of the context in which it is uttered. This also holds for sentences presented out of context, where interpreters presuppose certain background assumptions against which they interpret such sentences.
As regards the assumption that non-literal meaning is non-conventional and secondary (to literal meaning) in processing, arguments were given that this characterisation is not appropriate. Thus, given that conventionality is correlated with the speed of processing in interpretation and that this in turn can be measured by reaction times, numerous empirical studies indicate that non-literal meaning is not necessarily non-conventional and is not necessarily processed only after the literal meaning of the respective expression has been. In fact, speed of processing of – and thus the effort spent in interpreting – a particular meaning rather depends on that meaning’s salience, where this cross-cuts the literal/non-literal divide.
Such a view of literal meaning and non-literal meaning has consequences for the assumptions about the nature of lexical meaning. That is, if a simple expression’s literal meaning actually is not context-independent but one still wants to assume that semantic information stored in the lexicon IS context-independent, then an expression’s lexical meaning cannot be its literal meaning, but has to be something more abstract. In fact, arguments were presented in favour of a view of lexical meaning as being highly underspecified. Moreover, this view was shown to be best compatible with the revised characterisations of literal and non-literal meaning, as the minimally specified semantic forms of expressions in principle allow for both literal as well as non-literal specifications.
A more general conclusion drawn from the discussion of the standard notions of literal meaning and non-literal meaning is that since they do not adequately reflect the properties these two types of meaning actually have, the terms therefore cannot be used as basic concepts in definitions of semantics and pragmatics. In addition, the discussion showed that, actually, literal meaning and non-literal meaning are not so different from one another as traditionally assumed and therefore should not be characterised in terms of the dichotomies traditionally used. Moreover, since both terms refer to rather intuitive notions, one should actually not expect that a characterisation that captures those intuitions is precise enough for the two terms to be useful in distinguishing particular levels of meaning. This expectation is born out by the alternative characterisation of literal meaning and non-literal meaning I proposed in chapter 5.
Thus, I characterised literal meaning as basic and non-derived and non-literal meaning as non-basic, deviating or derived from some underlying basic meaning and in some sense replacing it. This rather vague characterisation captures metaphor and metonymy as prototypical cases of non-literal meaning, but also idioms and the results of ad-hoc concept formation. In contrast, such pragmatically determined meaning aspects as conversational implicature, indirect speech acts and the results of free enrichment are not captured by either of the characterisations as these types of meaning aspects are neither basic nor do they – in any intuitive sense – replace the meaning on which their determination is based.
From the discussion of the various approaches to the differentiation of semantics from pragmatics, the conclusion emerged that actually, none of the dichotomies traditionally used in the characterisation of the semantics/pragmatics distinction is in fact appropriate for this task. Thus, the pair of terms literal meaning vs. non-literal meaning in their standard characterisation cannot be used as it is inadequate. However, even with a more adequate characterisation these terms do not refer to meaning aspects that can be considered to result from distinctive processes and as such allow for a classification of being semantic or pragmatic in nature. The terms conventional meaning vs. non-conventional meaning cannot be used either, since, as we saw, their application presupposes that conventionality of meaning is an absolutive property, which, I argued, it is not. This leaves us with the dichotomy context-independent vs. context-dependent meaning, of which I argued that it can only be used to differentiate linguistic semantics from pragmatics. This is so because the process of semantic interpretation – traditionally assumed to be part of linguistic semantics – actually can only apply once the output of the context-independent linguistic semantics component has been pragmatically enriched to a proposition. However, the property of context-(in)dependence cannot be used to differentiate between the meaning aspects pragmatics as compared to real semantics deals with. This is so because pragmatics successively determines context-dependent meaning aspects for integration into the semantic form delivered by linguistic semantics and – at various points during the interpretation process – real semantics interprets those increasingly enriched forms. Thus, the two systems are actually differentiated by the nature of the processes that constitute them: monotonic reasoning with non-defeasible output in the case of real semantics vs. non-monotonic reasoning with defeasible output in the case of pragmatics.
As concerns the nature of the processes determining the various meaning levels identified in the approaches reviewed: whereas expression meaning is solely determined by the processes within linguistic semantics, which combine the underspecified lexical meanings of the expressions involved in accordance with the principle of compositionality, the levels of meaning what is said/utterance meaning and what is meant/communicative sense both result from enrichments of the thus determined expression meaning by particular pragmatic processes. However, even though both what is said/utterance meaning and what is meant/communicative sense are thus context-dependent levels of meaning, I argued that a differentiation between them should be made nevertheless. In particular, I claimed that the differentiating property of the processes involved in determining what is said/utterance meaning in contrast to those involved in determining what is meant/communicative sense is that the latter, but not the former, take into account assumptions concerning the speaker’s intentions in making the particular utterance. The claim that the process of free enrichment – taken to contribute to the level of what is said/utterance meaning – actually depends on a consideration of speaker intentions was refuted by assuming that during the interpretation of an utterance, information stored in conceptual frames is made use of and integrated into the semantic form built up for that utterance. With such an assumption the fact can be explained that the implicit meaning aspects (IMAs) taken to be contributed by free enrichment seem to even arise when they are not speaker-intended. This is because, for the determination of such IMAs, the speaker’s intentions in fact are not considered. Moreover, this assumption is also supported by the fact that such recourse to conceptual knowledge organised in frames seems to be necessary in any event for the establishment of the discourse relations holding between the individual utterances in discourses/texts.