Theoretical Problems in Lexical Field Analysis

Table of Contents

1 Introduction

2 Structuralism
2.1 The Saussurean Dichotomies
2.2 Componential Analysis
2.3 Sense Relations

3 Relativism and Functionalism

4 The Development of Lexical Field Theory
4.1 Trier and Weisgerber
4.2 Porzig
4.3 Coseriu

5 Folk Taxonomies

6 Criticism

7 Conclusion

8 Works Cited

1 Introduction

Asphyxiate, choke, drown, stifle, suffocate. One does not need to be trained linguistically to notice that most lexical items can be collected with other lexemes in order to form groups. Structuralist approaches in particular, but also semantic theories of different orientations, have come to investigate this phenomenon.

Semantics, the theory of linguistic meaning, is a broad field which touches on aspects of form and function, as well as on issues connected with a variety of areas including psychology, anthropology and lexicology. The area that is the focus of this paper is the concept of the lexical field. Originating with German and Swiss structuralists in the early 20th century, the fundamental notion behind this idea is the systematic nature of language. Since its first appearance in 1924 as ‘Bedeutungsfeld’ (literally field of meaning, cp. Geckeler 1971: 86-89), it has undergone a variety of changes in conception and terminology. These are in part due to a natural process of differentiation in terminology over time, but also to the difficulty of translating the originally German nomenclature.

In this paper, the term lexical field (‘Wortfeld’) will be applied in its broadest sense, that is, “a group of words closely related in meaning, often subsumed under a general term” (Lehrer 1974: 1). Competing terminology, such as semantic field (‘Bedeutungsfeld’) and conceptual field (‘Begriffsfeld’) will only be used in connection with the description of research history. The differentiation made by Lipka (1980) between lexical field and word-field, the latter denoting a subcategory of the former, will not be made.

While it has often been criticized, the notion of the lexical field can constitute a powerful tool in both intralingual and interlingual analysis. This paper provides a broad overview of topics which are relevant in the discussion of lexical field analysis. Afterwards, these topics will be brought into context by presenting a selection of outstanding approaches to field theory. Finally, the German word field tradition will briefly be compared to the parallel development in the USA.

2 Structuralism

As pointed out above, the analysis of lexical fields originates in the structuralist tradition. This chapter serves to delineate the relevant structuralist ideas in the context of lexical field study (cp. Löbner 2002: 127-130).

While the nineteenth century was dominated by historicism, the turn of the century saw a paradigm shift usually attributed to Ferdinand de Saussure and the publication of his Cours de linguistique Générale in 1916. His structuralism had an enormous influence on following linguistic study, both as a starting point for subsequent research (in Europe most importantly Jakobson and Trubetzkoy of the Prague School as well as Hjelmslev of the Copenhagen School, Bloomfield in the US) and as a trigger for a counter movement (Chomsky’s Formalism). Although structuralist ideas might not be irrefutable (cp. for example cognitive linguistics), they nevertheless form an important part of everyday linguistic life and provide a basis for lexical field analysis.

The central thesis of structuralism is that the individual units in a language system are interrelated and obtain their meaning, or value (‘valeur’), only by their relationship with and difference from the other units.^[1] This entails that, as soon as one unit changes its value, the whole system is affected. This approach can be applied to phonology, grammar and lexical structure. That the vocabulary of a given language is structured in a certain way becomes obvious when analyzed contrastively. For instance, the English homonym mole contrasts with German Maulwurf, Muttermal and Mole. (cp. Lyons 1977: 231-238)

2.1 The Saussurean Dichotomies

Saussurean structuralism premises on a set of four dichotomies. These have been of great significance in the development of structuralism and linguistics in general.

Firstly, language is regarded as a complex, abstract language-system of relations and rules (‘langue’) that forms the basis of all regularities in actual language use in a particular language-community: language-behavior (‘parole’). Saussure’s notion of the langue is ambiguous in that he underlines both its supra-individual und social character and its psychological validity. Furthermore, one may argue about the abstract and idealized assumption of an underlying, relatively uniform language system. (cp. Lyons 1977: 239)

The second distinction refers to the linguistic sign, which consists of two parts, signifier and signified, or structure and substance (‘signifiant’ and ‘signifié’), which are associated arbitrarily (cp. Lyons 1977: 239-241). It is not inherent in the concept kill someone deliberately that it must be expressed by the lexeme ‘murder’; the lexeme ‘bread’ would serve equally well. This differentiation between signifier and signified is implicit in lexical field analysis, and manifests itself in the question of whether one’s approach should be semasiological or onomasiological; that is, starting from the form and asking for the meaning, or vice versa.

Thirdly, the units of a system can be combined into more complex constructions, or syntagms, with other units of the same level and according to language-specific formation rules. For each position in such a complex unit, the speaker chooses between different options in a set of alternatives, the paradigm. Saussure’s distinction between syntagmatic and paradigmatic (in his original terminology: associative) relations is best illustrated with an example:

Abbildung in dieser Leseprobe nicht enthalten

If one assumes that these complementary concepts of paradigmatic selection and syntagmatic combination constitute the structure of a language-system at every level it follows that, in order to describe a lexical field satisfyingly, one has to consider both axes of structure (cp. ch. 2.3). (cp. Lyons 1977: 241-243)

The fourth Saussurean dichotomy is that of synchrony and diachrony, the study of a language at a certain point in time and the investigation of changes in a language between two points in time. Naturally, the term synchronic language-system is a theoretical construct, as two dialects may for example differ more from each other than two stages in the development of one dialect. However, it cannot be argued that the distinction between diachrony and synchrony is methodologically essential. (cp. Lyons 1977: 243-245)

2.2 Componential Analysis

One method which has been developed in the framework of structural semantics is componential analysis. As a tool to discover, predict and explain the semantic relations among the items in a lexical field, it is an attempt to provide field theory with a solid theoretical and methodological basis (cp. Storjohann 2003: 30). The crucial insight of componential analysis is that word meanings are made up of atomic elements or components (cp. Evans 2011: 919). This notion of distinctive features goes back to the oppositional character of linguistic signs according to Saussure. As was introduced above, in traditional structuralism a sign’s meaning is thought to be exclusively constituted by its difference to the other signs in a language system. This idea was adapted to phonology by Prague School linguist Trubetzkoy, who describes phonemes according to a set of binary, phonological features such as +/- voice. (cp. Storjohann 2003: 30)

Trubetzkoy’s colleague Jakobson as well as Hjelmslev (1974 [1943]) of the Copenhagen School and Coseriu (1966), amongst others, introduced the idea of distinctive features into lexical semantics (cp. Coseriu and Geckeler 1981: 42). In componential analysis, lexical meaning is examined by decomposing the content of a linguistic sign into smaller elements, or semantic features (cp. Löbner 2002: 132-142). The values of the entirety of these semantic features and the means by which they are combined are thought to add up to the value of a lexeme. According to Hjelmslev (1974 [1943]), one has to trace back the open class of lexical units to closed paradigms. Within these, one can then without effort determine minimal changes on the signified side of a sign which lead to alterations of the signifier. For instance, if the component female is replaced by its counterpart male in the lexeme mother, the result is a change from mother to father. (cp. Storjohann 2003: 30)

An especially intriguing possibility of componential analysis is the potential to determine universal meaning components across different languages. Many believe that the elementary components of sound and meaning are language-neutral, while complexes of components such as phonemes and word-meanings as well as their paradigmatic and syntagmatic relations are unique to particular languages. Theoretically, one has access to a set of potential distinctions, a subset of which is realized in each language (cp. Saussure’s thesis of the continuity of substance). (cp. Lyons 1977: 245-246)

One such approach is that by Polish-Australian linguist Anna Wierzbicka, who has undertaken to identify semantic primitives or primes, indefinable words which serve to define everything else (cp. 1996: 9-23). Since the 1970s, she has been able to establish a set of 55 primitives such as you, kind, and part. She argues that, without primitives, every description of meaning will ultimately be circular and thus untenable (as when, for example, cranky is defined by grouchy, which in turn is illustrated by crotchety, which is explained with cranky). Wierzbicka furthermore proposes that the combination of the shared core of all natural languages forms a natural semantic metalanguage (NSM). Using s emantic explications in the NSM, one can paraphrase any word in any language. This type of description will often turn out unidiomatic, but still intelligible and, very importantly, free of any semantic re-coloring. As this approach certainly seems rather abstract, an example is appropriate. The following semantic explications describe the verb lie:

someone X lied to someone Y:
someone X said something to someone else Y
this someone knew that it was not true
this someone said it because he/she wanted this other someone to think that it was true
people think that it is bad if someone does something like this^[2]

While the notion of semantic primitives has often been dismissed as arbitrary, stating that primitives could be broken down further by means of another system (cp. Wierzbicka 1996: 11), it certainly poses an interesting approach. It should be mentioned that a number of scholars regard semantic relations as primitives in the sense that these cannot be further defined or subdivided (cp. Bolin 2005: 13).

Another approach incorporating componential analysis is that by Katz and Fodor, which resulted in the integration of semantics and syntax into transformational grammar (cp. Lyons 1977: 318). Katz and Fodor (1963: 186-187) introduce a dichotomic distinction: those elements which express features relevant in a language-system, for instance female, are called semantic markers. In a dictionary entry, these reflect systematic semantic relations which hold between that item and the rest of the lexicon. Semantic distinguishers, on the other hand, are used for any further semantic characteristics, reflecting what is idiosyncratic about its meaning. This distinction was extensively criticized and eventually abandoned in favor of a onefold set of semantic components (cp. Lehrer 1974: 50).

2.3 Sense Relations

Although the structuralist view that the description of meaning relations completely exhausts the description of meaning can be argued with, one cannot refute the importance of meaning relations in semantic analysis.

In lexical field theory, the study of syntagmatic relations may offer valuable insights. For instance, while the German lexeme ersticken unequivocally includes the victim’s death, in English this may only be expressed unambiguously on the syntagmatic level, either with the collocation choke sb. to death or in the larger context, such as this instance found in an online forum:

That's what I'm thinking. If he would have "choke" choked him(choked with the intent to kill), the buy would've been pronounced dead at the scene or shortly there after, not 45 minutes later.. [sic!]^[3]

Paradigmatic meaning relations, denoting the relations of sense which hold between the members of a lexical field, can be considered more important in lexical field analysis. One may be content with a simple fourfold grouping into synonymy (suffocate/asphyxiate), hyponymy (kill/drown), meronymy (drown/die)^[4] and opposition (kill/re­surrect). However, a more finely granulated approach can be advantageous. Lyons (1977: 270-335) gives a comprehensive account of the basic principles of the theory of semantic fields in terms of sense-relations which can only be roughly sketched in the following section.

Trier, whose pioneering approach will be discussed below, opens his major work with the words “Jedes ausgesprochene Wort läßt seinen Gegensinn anklingen”^[5] (1931: 1). Lyons (270-292), too, begins with a discussion of lexical opposition. Criticizing precedent approaches and incompatibilities with, for instance, logic, he proposes the following sub­cate­go­ri­zation (cp figure 2 below): Contrast can be regarded as an umbrella term which does not imply the number of elements in the set, while opposition is by definition binary. Antonymy, in turn, denotes gradable opposites, and ungradable opposites are called complementaries. A further distinction on this level is converseness, that is, pairs of lexemes which express the same relation with reversed roles. A fourth type is directional opposition, which denotes opposite directions on an axis with respect to a given place P. Contrasting with opposition, there are also various

[...]

^[1] Note the difference to cognitive approaches, which assume that the meaning of a lexeme can in principle be studied individually, not as part of a linguistic but of our overall cognitive system. (cp. Löbner 2002: 128)

^[2] Taken from the Natural Semantics Metalanguage homepage http://www.une.edu.au/bcss/linguistics/nsm/explications.php. Accessed March 28, 2012.

^[3] http://www.camaroz28.com/forums/f-body-lounge-24/shoplifter-dies-workers-chokehold-752291/page8/

Accessed 05.04.2012.

^[4] Cp. Maziarz e.a.’s (2011: 191-192) notion of a verb relation of meronymy based on the system of noun meronymy and holonymy.

^[5] Each pronounced word calls forth its opposite. (LRK)