Chapter 1 – Language documentation: What is it and what is it good for?
19
b.
In line with the structuralist conception of the language system, gram-
mars and dictionaries contain abstractions based on a variety of analyti-
cal procedures. With the data contained in grammars and dictionaries,
most aspects of the analyses underlying the abstractions are not verifi-
able or replicable. There is no way of knowing whether fundamental
mistakes have been made unless the primary data on which the analyses
build are made available in toto as well.
c.
Grammars usually only contain statements on grammatical topics which
are known and reasonably well understood at the time of writing the
grammar. Thus, for example, grammars written before the advent of
modern syntactic theories generally do not contain any statements re-
garding control phenomena in complex sentences. Many topics of cur-
rent concern such as information structure (topic, focus) or the syntax
and semantics of adverbials have often been omitted from descriptive
grammars due to the lack of an adequate descriptive framework. As
pointed out in particular by Andrew Pawley (1985, 1993, and elsewhere),
there is a large variety of linguistic structures often subsumed under the
heading of speech formulas which do not really fit the structuralist idea
of a clean divide between grammar and dictionary and thus more often
than not are not adequately documented in these formats.
d.
Grammars and (to a lesser extent) dictionaries provide little that is of
direct use to non-linguists, including the speech community, educators,
and researchers in other disciplines (history, anthropology, etc.).
These points of critique mostly pertain to the fact that structuralist language
descriptions are reductionist with regard to the primary data on which they
are based and do not provide access to them. Or, to put it in a slightly dif-
ferent and more general perspective, they document a language only in one
of the many senses of “language”, i.e. language as an abstract system of
rules and oppositions. Inasmuch as structuralist language descriptions are
intended to achieve just that, the above “critique” is, with the possible ex-
ception of point (b), not fair in that it targets goals for which these descrip-
tions were not intended.
12
In this regard, it should be emphasized that the above
points in no way
question the usefulness and relevance of descriptive grammars and diction-
aries with regard to their main purpose, i.e. to provide a description and
documentation of a language system. While there is always room for im-
provement (compare points (b) and (c) above), there is no doubt about the
fact that grammars and dictionaries are essentially successful in delivering
20
Nikolaus P. Himmelmann
system descriptions. What is more, the above points also do not imply that
grammars and dictionaries do not have a role to play in language documen-
tations, as further discussed in the next section. The major thrust of the
critical observations above is that a description of the language system as
found in grammars and dictionaries by itself is not good enough as a lasting
record of a language, even if accompanied by a text collection. And it is
probably fair to say that the way primary data have been handled in the
grammar-dictionary format is now widely seen as not adequate and thus in
need of improvement.
From this assessment, however, it does not necessarily follow that the
basic format of Table 1 is the only imaginable format for lasting, multipur-
pose records of a language. Instead, it may reasonably be asked, why not
combine the strong sides of the two formats discussed so far and propose
that language documentations consist of the combination of a large corpus
of annotated primary data as well as a full descriptive grammar and a com-
prehensive dictionary? This is the question to be addressed in the next sec-
tion.
4.2. An extended format for language documentations
Assuming that the structuralist notion of a language as a system of rules
and oppositions is a viable and useful notion of “a language”, though not
necessarily the only useful and viable one for documentary purposes, and
assuming further that a descriptive grammar and a dictionary provide ade-
quate representations of this system, it would seem to follow that a truly
comprehensive language documentation does not simply consist of a large
corpus of annotated primary data – as sketched in Section 3 – but instead
should also include a comprehensive grammar and dictionary. Along the
same lines, one may ask why the apparatus in Table 1 should only contain a
sketch grammar and not a fully worked-out comprehensive grammar, thus
replacing the format in Table 1 with the one in Table 2.
13
Chapter 1 – Language documentation: What is it and what is it good for?
21
Table 2. Extended format for a language documentation
Primary data
Apparatus
Per session
For documentation as a whole
Metadata
General access resources
–
introduction
–
orthographical conventions
–
glossing conventions
–
indices
–
links to other resources
…
recordi
n
gs
/r
ecords
o
f obs
er
va
bl
e l
ingui
st
ic
behavi
or
and
met
al
ingui
st
ic
k
now
le
dge
Metadata
Annotations
–
transcription
–
translation
–
further linguistic and
ethnographic glossing
and commentary
Descriptive analysis
–
ethnography
–
descriptive grammar
–
dictionary
The difference between the basic format for language documentations in
Table 1 and the extended format depicted in Table 2 pertains to the addition
of fully worked out descriptive analyses on various levels (as indicated by
the shaded area in Table 2), replacing the corresponding sketch formats
(sketch grammar, ethnographic sketch) under general access resources in
the basic format. Whether this is in fact a fundamental difference or rather a
gradual difference in emphasis, is a matter for further debate. In actual
practice, the difference may not be as relevant as it may appear at first sight,
as we will see at the end of this section. Still, in the interest of making clear
what is involved here, it will be useful to highlight the differences between
the two formats and to indicate some of the problems that are created by
incorporating comprehensive descriptive formats in the extended documen-
tary format. There are at least two types of such problems, one relating to
theoretical issues, the other to research economy.
The theoretical problem pertains to the fact that it is not at all clear how
exactly the descriptive grammar (or the ethnography or the dictionary)
14
should look that is to be regarded as an essential part of a language docu-
mentation. As is well known, for much studied languages such as English,