How conceptual are semantic maps?

The question addressed in this paper is whether (and to what extent) a semantic map aimed at representing the multifunctionality of a given construction (or set of constructions) in discourse can be thought of as endowed with conceptual reality. To be considered as a mental representation that is essentially similar in all human brains, such a map should meet two requirements: (i) its nodes should be bundles of semantic and pragmatic properties that form conceptual archetypes, that is, ways of conceptualizing and categorizing dynamic or static configurations that are fundamental to human experience; (ii) there should be a high degree of regularity in the data material, i.e. each construction should be associated with a node or a contiguous set of nodes in a regular way. However, observing the use of grammatical constructions in discourse provides us with compelling evidence that discourse contexts are complex entities involving many different variables, and that “a perfect fit is not the usual state of affairs for models of complex human behavior (including language)” (Croft and Poole 2008: 6). Based on a previous analysis of various passive and impersonal constructions in a parallel corpus of five European languages, I will argue that a first-generation semantic map representing the distribution of these constructions in discourse and comprising a few conceptual archetypes may be only an idealized abstraction over the conflicting evidence of the association between discourse contexts and construction types. As an idealization, such a map is not particularly informative as to language-specific tendencies and idiosyncracies, and does not allow us to analyze all the datasets that we might be interested in analyzing. On the other hand, a second-generation semantic map proves to be a more reliable tool for representing variation in discourse and does not force the analyst to posit (and multiply) conceptual structures where there may be none. 1. Semantic maps and conceptual reality Some practitioners of the semantic-map method are rather reluctant to consider semantic maps as “mental” or “cognitive” maps, i.e. as “direct representation[s] of the relationships between meanings in speakers’ minds” (Haspelmath 2003: 233, my emphasis). According to others (e.g. Croft 2001), semantic maps make conceptual sense and provide objective evidence for which meanings are cognitively closely related. In Croft’s words, they accurately reflect “the geography of the human mind, which can be read in the facts of the world’s languages in a way that the most advanced brain scanning techniques cannot even offer us” (Croft 2001: 364). Representatives of the former approach usually maintain a neutral stance as to whether nodes on semantic maps represent different uses (= contextual meanings) or different senses (= conventional meanings) of grams (e.g. Haspelmath 2003: 212-213). The essential idea in the latter approach is that “the use of a single grammatical form ... for a set of functions implies that speakers conceptualize those functions as similar or related to one another” (Croft and Poole 2008: 5). Those who credit conceptual reality to semantic maps generally assume that nodes on semantic maps are conceptual struc1 I wish to thank Martin Haspelmath and an anonymous referee for their insightful comments on a first draft of this paper. tures or conceptual archetypes (Kemmer 2003: 98); that is, they represent ways of conceptualizing and categorizing dynamic or static configurations that are fundamental to human experience (Croft 2001: 98; Kemmer 1993, 2003: 97ff.). The arrangement of nodes on a map mirrors the arrangement of the corresponding conceptual structures (or functions) in the speaker’s mind (Croft 2001: 93). Kemmer’s statements in (1)-(2) should serve as an illustration of this position: (1) Situation types [i.e. nodes on the semantic map of event elaboration, AS] can be thought of as sets of situational or semantic/pragmatic contexts that are systematically associated with a particular form of expression. By ‘semantic/pragmatic contexts’ I do not mean simple ‘real world contexts’ existing independently of the language-user; situational contexts include ‘real world’ information, but that information is necessarily filtered through the conceptual apparatus of the speaker. This conception of situational contexts thus allows for the obvious role of the language-user in construing particular real world situations in different ways (Kemmer 1993: 7, my emphasis) (2) Marking systems in the reflexive/middle domain integrally involve alternative conceptualizations of participant structure and event structure as a whole. Languages may differ as to the morphosyntactic means they have conventionalized for coding such differences in conceptualization, but the variation is highly constrained by the underlying conceptual system (Kemmer 2003: 115). Conceptual structures are not intended as merely semantic in the truth-conditional sense of this term. Rather, they are meant to represent all of the conventionalized knowledge associated with a given gram, which may also include the discourse conditions under which the gram is used (Croft 2001: 19). This position rests on the (more or less implicit) assumption that there is no separate pragmatic component, and that any linguistic property traditionally analyzed in terms of pragmatics is semantically or conceptually determined in the first place. As Croft (2001: 93) puts it, “conceptual space also represents conventional pragmatic or discourse-functional or information-structural or even stylistic and social dimensions of the use of a grammatical form or construction”. As a consequence, semantic maps lend themselves well to being applied to phenomena that are sensitive to discourse conditions such as the relative prominence and the information status of arguments and situations (e.g. voice constructions, Croft 2001: 283ff.; the coordination/subordination continuum, Croft 2001: 320ff.). The question addressed in this paper is whether (and to what extent) a semantic map aimed at representing the multifunctionality of a given construction (or set of constructions) in discourse can be thought of as endowed with conceptual reality. To be considered as a mental representation that is essentially similar in all human brains, such a map should meet two requirements: (i) conceptual relevance: its nodes should be clearly identifiable bundles of semantic and pragmatic properties that form conceptual archetypes in the sense of Kemmer (2003: 98); (ii) high degree of regularity in the data material: these bundles of features should represent the typical contexts of use of a given (set of) construction(s), and each construction should be associated with a node or a contiguous set of nodes in a regular way. However, observing the use of grammatical constructions in discourse – their “natural habitat” (Givón 1995: 309) – provides us with compelling evidence that discourse contexts are complex entities involving many different variables, and that “a perfect fit is not the usual state of affairs for models of complex human behavior (including language)” (Croft and Poole 2008: 6). A certain amount of “noise” remains: what we usually find are preferences and tendencies, not categorical form-function mappings. Occasionally, forms are mapped onto functions that are generally paired with other forms in a given language. The noise is particularly rampant when searching for functional equivalents across languages (for instance looking at translation equivalents in parallel texts).