![Sample our Language & Literature journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5941/TOC-Subject-Sample-2021-Language-Literature.jpg"})
Abstract
This paper discusses the relevance of the distinctions made by Hjelmslev ([1943] 1961), between purport, substance and structure on the one hand, and content and expression on the other, for the theory of Functional Discourse Grammar (FDG). It is shown that FDG, as a form-oriented theory, shares a number of aspects with Hjelmslev’s model; at the same time, however, it is argued that, as a function-to-form theory, FDG assigns a much more important role to the notion of substance. Three major differences between the two approaches are identified, concerning (i) directionality in the model (between substance and structure), (ii) motivation (of structure on the basis of substance) and (iii) the difference between “structure” (in the form of more or less stable primitives) and “structuring” (as a process). Subsequently, a corpus-based, diachronic analysis of the use, form and development of a number of related binominal noun phrases with the head noun beast (e.g. a beast of a party) is offered to illustrate the FDG interpretation of the distinction between substance and structure, and to justify the greater importance attributed to the notion of substance. It is concluded that compared to Hjelmslev’s model, FDG is better equipped, both diachronically and synchronically, to deal with the specific features of and relations between the four types of binominal construction discussed in the paper.
Notes
1 The term construction is used here and in the rest of the paper in a theory-neutral way. Constructions as defined in Construction Grammar are not part of the FDG framework; instead FDG makes use of interpersonal and representational frames (to capture certain pragmatic and semantic aspects of an utterance) and morphosyntactic and phonological templates (to code this pragmatic and semantic information).
2 In this respect, FDG differs from its predecessor, Dik’s Functional Grammar (e.g., Dik Citation1997a, 1997b). Dik’s model (in particular its earlier versions [e.g., Citation1978]) had a much clearer semantic focus (influenced by predicate logic), in which predicates were seen as the basic building blocks of linguistic expressions.
3 For a (much more detailed) discussion of the relation between Hjelmslev’s distinctions and the several stages of Hallidayan Systemic Functional Grammar (originally based on Hjelmslev’s model), see Taverniers (Citation2011, Citationforthcoming).
4 Hjelmslev ([Citation1943] Citation1961, 76) later dismisses the idea of purport as what is common to all languages as an illusion, since “the purport is formed in a specific fashion in each language, and therefore no universal formation is found.”
5 ‘Expression-form’ belongs to the realm of phonology: it refers to the interpretation of a sound chain in terms of the phonological system of a specific language. ‘Content-form’ is a much more elusive concept, in particular because Hjelmslev himself does not provide a clear definition or example. As pointed out by Taverniers (Citation2011, 1105), “content-form is not linked by Hjelmslev to a specific component of linguistics, and it has been interpreted in different ways by linguists ever since it was introduced.”
6 Note furthermore that in FDG, the communicative intentions included in the Conceptual Component are potentially universal, i.e., not yet formed by a specific language. This seems to be compatible with Hjelmslev’s definition of substance as what is structured by the language system. Hjelmslev does not, however, explicitly discuss the (universal) nature of purport or substance itself; like Saussure, he is interested only in the formal similarities and the differences between languages (e.g., Hjelmslev [Citation1943] Citation1961, 76).
7 The main criterion for distinguishing between lexical and grammatical (functional) information in FDG is that of modifiability: if an element allows modification, it is analyzed as lexical, if not, it is analyzed as grammatical (Hengeveld and Mackenzie Citation2008, 59). Elements with an in-between status tend to exhibit non-default behavior, in particular in the area of modification (as shown in Section 3.2). This does not mean that other factors (such as the possibility of Focus assignment) cannot play a role; as pointed out by Keizer (Citation2007a), it is exactly the presence of a number of factors that is responsible for the gradient nature of the distinction. On the other hand, it can be argued that, faced with a choice, speakers have to choose how to code a particular bit of information, which means that in many cases even the in-between instances are linguistically coded as either lexical or grammatical.
8 There are, of course, exceptions, as in the case of iconic relations.
9 Here, too, there are exceptions, since in some cases ‘a-functional’ factors may play a role (e.g., structural complexity or priming).
10 Some changes are not functionally motivated (by commucative need), but rather by ease-of-effort (e.g., phonological changes). Such changes are, of course, not triggered at the conceptual level. The particular development discussed in this paper actually involves both types of changes.
11 It is important to realize that all information represented in the model is linguistic. At the Interpersonal Level, the various units represent strategic acts on the part of the speaker; at the Representational Level, they represent the kind of entity designated, not the entity itself. Semantic categories are thus linguistic categories, representing those ontological distinctions that are relevant for the analysis of a particular language (Hengeveld and Mackenzie Citation2008, 130–131).
12 Placeholders are used for those elements whose final form can only be determined at the Phonological Level, i.e., for allomorphs.
13 The EBNP (a term taken from Trousdale [Citation2012, 179]) is the same phrase that Aarts (Citation1998) and Keizer (Citation2007b, 85–106) call the BNP, Foolen (Citation2004) the expressive BNP, and Ross (Citation1973) and McCawley (Citation1987) an adjective noun.
14 It would seem reasonable to assume that beast would also be found in the possessive construction, but there is no historical evidence to support this intuition.
15 This incongruity between semantic function and syntactic form has been remarked upon by Ross (Citation1973) and McCawley (Citation1987) in a Generative Grammar analysis and who included EBNPs in their Adjective–Noun category, “[which] mimic noun syntax with adjective syntax” (McCawley Citation1987, 459).
16 Trousdale (Citation2012, 182–183) also discusses this development, and he claims that it is a reanalysis of the degree modifier form of ‘hell-of,’ the next construction on the grammaticalization cline. Due to space constraints, this question cannot be discussed in detail; however, we would like to point out that he does not make the EBNP and evaluative modifier distinction, which is a critical division.