Chapter 6
Documenting lexical knowledge
John B. Haviland
Introduction
Lexicography, the practice of documenting the meanings and uses of
“words” (literally by “writing” them down), is, through its products, per-
haps the most familiar branch of linguistics to the general public. It is also
an ancient and much theorized activity. In the Boasian trilogy for language
description of grammar, wordlist, and text, it is surely the dictionary whose
compilation is most daunting. The process begins with a learner’s first en-
counters with a language, and it ends, seemingly, never. Worse, it is an en-
deavor fraught with doubt, centrally about when enough is enough both for
the whole – when one should assume that the basic or most common words
of a linguistic variety have been captured and characterized – but also for
any single putative dictionary entry, given the apparent endless variety of
nuance and scope for words and forms, not to mention the idiosyncrasies of
compound or derived expressions. Moreover, despite bounteous speculation,
from many disparate linguistic traditions, on what metasemantic devices
one might employ to capture meanings, despite multiple models and exam-
ples of the results of dictionary-making, and despite ample experience, for
most of us, in the ordinary business of “explaining the meanings of words,”
doubt is likely to assail us on every single effort: have we said enough?
have we forgotten something? did we get even this single word right?
This chapter introduces techniques and concepts relevant to producing a
lexical database as part of a language documentation project. I concentrate
on a series of doubt-producing obstacles for the field lexicographer, with
some suggestions about how at least to address, if not to overcome them.
My coverage is deliberately partial. I draw heavily on my own fieldwork in
Mexico and Australia, to consider three general issues. First, I review fa-
miliar morals about the nature of word meaning – concepts from linguistic
philosophy that are easy to forget in the heat of the lexicographic moment.
Second I consider semantic metalanguages proposed to deal with different
kinds of meaningful elements, from “functional” to lexical and from roots
130
John B. Haviland
to stems. Third, and most centrally, I review techniques for systematically
extracting lexical knowledge. I largely ignore several related and important
topics: lexical variation and how to represent it (see Chapter 5), ideological
issues inescapably involved in promulgating any dictionary (see again
Chapter 5, and the discussions in Frawley et al. 2002), and wider issues in
lexical semantic theory (about sense relations, problems of extension vs.
intension, etc.), which underlie all lexicographic practice but are beyond the
present scope. I begin with a highly selective review of published materials
on lexical knowledge, especially as relevant to documenting endangered
languages.
1. Lexicography and its products
In addition to a large theoretical literature on meaning, there is a practical
tradition of dictionary-making that has spawned handbooks and histories,
as well as essays on the lexicographer’s craft. These rarely provide solace
for the field worker.
The lexicon, in modern linguistics, has come to mean a repository for
otherwise anarchic facts, an inventory of arbitrary pairings of pronuncia-
tions with bundles of features. It is where language stores its idiosyncrasies
and irregularities. What systematicity there is to the lexicon so conceived
derives from feature systems themselves, taken to represent syntactic and
semantic patterning underlying surface lexical forms. Studying such pat-
terning is the usual province of lexical semantics, which catalogues various
relations between the senses of members of different subsets of lexical
forms (Cruse 1986), systematic properties of surface word classes or “parts
of speech,” facts of argument structure, diathesis, and the like. The main
contribution to linguistic theory of much empirical lexicography has been
in elucidating semantic and syntactic interrelationships at the level of the
surface word (Levin 1993).
Field linguistics, once the province of anthropological linguists, gave rise
to much of the underlying conceptual apparatus of lexical semantics. Early
theories pursued an analogy between phonological features and the “com-
ponents” of meaning in structured sets of “folk terminology,” from kinship
to ethnobotany, from pronoun systems to verbal typologies. The classic
studies of “ethnoscience” investigated culturally elaborated lexical systems,
particularly in “natural” domains like ethnobotany. Further empirical inspi-
ration for semantic theorizing came, for example, from the languages of
Chapter 6 – Documenting lexical knowledge
131
Aboriginal Australians, celebrated for their linguistic acuity and creative
genius. Dyirbal verb semantics and the properties of special Dyirbal
“mother-in-law” vocabulary for affinal avoidance led Dixon (1971) to pos-
tulate a fundamental difference between semantically basic or “nuclear”
words, requiring some sort of decomposition into sublexical meaningful
dimensions, and non-nuclear words which could be defined in terms of the
nuclear words plus other devices of the grammar. Verbal play in ritual lan-
guage games learned by Warlpiri and Lardil initiates suggested that Abo-
riginal ethnolinguists had developed sophisticated semantic analyses of
ordinary vocabulary (Hale 1971, 1982).
The classic reference manual on lexicography is Zgusta (1971).
1
Of special
interest to the field lexicographer is Frawley et al. (2002), a collection of
essays by practicing lexicographers working on American Indian lan-
guages, which also considers problems in creating a lexicographic practice
in communities without one.
2
These range over theoretical issues in lexical
semantics (the nature of definition, the range of lexical knowledge that
speakers possess or a dictionary might include, and the interplay between
diachronic and synchronic lexical facts); to questions of representational
form, to sociopolitical issues in dictionary making (for whom is a dictionary
compiled and for what purposes; or, what kinds of sociolinguistic catego-
ries – specialized speech genres, gender or class specific lexical forms, for
example – are to be distinguished). These works go well beyond the limited
selection of topics addressed here.
The field linguist need not be a semanticist, except “for practical pur-
poses,” and lexicography in the service of documentation needs to strike a
balance between opposing desiderata. For example, in what sense is “com-
pleteness” – however that might be defined for an endangered language –
something to strive for? What about the mix of theoretically versus practi-
cally motivated metalanguages for representing lexical information? In the
field one should avail oneself of all possible tricks: bilingual dictionaries,
for example, can often start with existing word lists, in either the source or
the target language, and there is no reason to stand behind strict methodo-
logical principles or purism in generating lexemes for incorporation into a
lexical database.
Different lexicographic products reflect different starting points and
goals for compilers of lexical databases. Zgusta (1971) dedicates separate
chapters to the distinct issues involved in compiling polylingual (usually
bilingual) versus monolingual dictionaries. The contrast, and the choice of
Dostları ilə paylaş: |