LINGUIST List 10.1439

Sun Oct 3 1999

Review: Donohue: Grammar of Tukang Besi

Editor for this issue: Andrew Carnie <>

What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Andrew Carnie at


  1. klamerm, review: Donohue: Grammar of Tukang Besi

Message 1: review: Donohue: Grammar of Tukang Besi

Date: Mon, 20 Sep 99 12:06:55 MET
From: klamerm <>
Subject: review: Donohue: Grammar of Tukang Besi

Mark Donohue, 1999, A Grammar of Tukang Besi [Mouton Grammar Library 20], 
Berlin/New York: Mouton de Gruyter, 576 p.

Reviewed by Marian Klamer, General Linguistics, Vrije Universiteit Amsterdam 
and Leiden University, The Netherlands.

Tukang Besi is an Austronesian (Western Malayo-Polynesian) language, spoken 
by approx. 80.000 speakers in Southeast Sulawesi, Indonesia. The language is 
spoken on a small archipelago with the same name, located in the Banda and 
Flores Sea and consisting of the four islands Wanci, Kaledupa, Tomea, and 
Binongko. Tukang Besi is an exonym from Malay 'tukang besi' "blacksmith": the 
wares of the blacksmiths of the poorest island of the four, Binongko, are 
known as far afield as Ujung Pandang in South Sulawesi, and the blacksmiths 
of Binongko are credited with secret powers that enables them to pull glowing 
iron from the hearth without thongs, and to beat metal into machetes with 
their hands if necessary.

"A Grammar of Tukang Besi" is the first comprehensive grammar of the 
language. It is based on primary data gathered by the author in cooperation 
with several native speakers during field work in the area. Apart from a 
475-page description of the structural aspects of the language, the book also 
contains 40 pages of glossed and translated texts, a word list of about 1000 
items (Tukang Besi - English / English - Tukang Besi) and a small index.


Chapter 1 presents information on the setting and region of Tukang Besi, on 
the language's dialects, the attitudes of its speakers, and the sources for 
the study. The grammar is based on Donohue's own fieldwork, the majority of 
which took place 'between 1992 and 1995', in 'the village of Patuno on the 
north coast of Wanci, and also a fair deal of time around Kota Wanci on the 
west coast.' The language helpers were from 'a wide cross-section of the 
Tukang Besi speaking community'. I take this to imply that the grammar is 
based on data that are not skewed towards the idiolect of one informant, and 
that the dialect described is (one of the) Wanci dialect(s) (in chapter 2 it 
is explicitly stated that it describes the phonology and morphophonology of 
the Wanci dialect). 

The structure of Tukang Besi is described working upwards from the segmental 
to the sentential level in chapter 2 to 20. Rather than summing up the 
content of these chapters, in what follows I will highlight some aspects of 
TB that may be of interest to typologists and/or theoreticians. 

The Tukang Besi (henceforth TB) consonant set includes two implosive stops 
/b, d/, and a set of prenasalised segments, including voiced and unvoiced 
stops /mp, mp, nt, nd, nk, ngg/, /ns/ and a loan affricate /ndz/. There are 
several morpho-phonological processes involving nasals: -um- infixation to 
realise irrealis mood; prefixation of hoN-, where the nasal substitutes the 
root initial consonant, taking over its place features; and in root/foot 
reduplication, an initial voiceless stop is prenasalised: notinti 'he is 
running' > notinti-ntinti 'he's running around madly' .

The major constituents of a TB clause are the verb phrase and the case phrase 
(an NP governed by a case marker). In a verbal clause, word order is 
verb-initial. The subject is marked by a prefix on the verb, the object by a 
verbal enclitic. Nominal constituents are optional if the referential 
identity of the arguments they express has already been established. If a 
clause contains NPs, the order is V O A, but this order is not fixed.

TB is a Philippine-type language like e.g. Tagalog, in the sense that it is 
unlike languages with predominantly accusative or ergative syntax: each 
clause must morphologically select one of the arguments of the transitive 
verb as its pivot, and there is no unmarked choice. The selected argument is 
assigned a specific case, regardless of whether it is in A, S, or O syntactic 
role. In Philippine studies, this argument has traditionally been referred to 
as the 'focus', 'subject' or 'topic'. As these terms carry other associations 
as well, Donohue uses the term 'nominative' to refer to the case assigned to 
it, which is expressed by the article na. The other core article te assigns 
non-nominative case. Transitive verbs with subject and object markers assign 
nominative case (na) to the O syntactic role and non-nominative case to A; 
schematically: s-V-o [na O] [te A]. Intransitive verbs assign nominative case 
to their single argument S: s-V [na S]. Interestingly, if a transitive verb 
is not crossreferenced for its object, the case assignment is reversed: s-V 
[te O] [na A]. Also, the word order is now rigidly [VO]A. 

The author analyses this pattern as a Phillipine-type voice system, with a 
restricted range of categories -- only two. In TB the diachronic shift 
towards head-marking pronominal indexing has proceeded to quite an extent, 
but at the same time the overt Phillipine-type case system has been 
preserved, and its verbal cues reinterpreted as being those involving the 
presence versus absence of the object agreement.
A TB verb phrase contains the verb and either a bound pronominal object or an 
object case phrase (KP), but not both. The arguments that may be indexed on 
underived verbs are limited to Agent, Dative, Instrument or Theme/Patients. 
Verbal indexing is only available for so-called 'core arguments'. Core 
arguments are obligatorily marked by the case-marking articles na 
(nominative) or te (non-nominative), while oblique arguments may drop the 
oblique article i. Only core arguments may launch floating quantifiers or 
adverbs (chapters 7 and 20), may be relativised (chapter 15), and may be 
marked on the verbs of subordinate clauses such as nominalisations (chapter 
12) and relativisations. 

Nominal constituents are discussed in chapter 5 (pronouns), 6 
(demonstratives), 12 (noun phrases: core and oblique), 13 (possession and 
possessive constructions), 15 (relative clauses), and 18 (conjoining). A TB 
noun phrase is head initial. Pronominal possessor marking appears on the head 
noun if the NP is nominative [Nom [[N-Poss] Adj]], while in non-nominative 
NPs, the possessor is marked on the nominal modifier [non-Nom [[N Adj] 
-Poss]]. An NP preceded by an article (which functions as the case marker) is 
nested within a case phrase, KP. Both NPs and KPs may be prepositional 

Tukang Besi has two major open word classes: nouns and verbs. Nouns 
characteristically appear inside NPs, are preceded by an article, a 
preposition or a numeral classifier. Verbs are the class of words that can be 
prefixed for subject when used as the head of a main clause. Adjectives are a 
subclass of the non-agentive intransitive verbs. The category of adverbs is 
not discussed in the chapter on word classes, but from the information given 
elsewhere in the book, it seems to be a separate category which shares some 
properties with the verbal class. 

In TB the same word form may be used, with no derivational morphology, in 
both nominal and verbal frames. In fact, such 'pre-categorial' items 
constitute the majority of the lexical items in the open word classes. This 
phenomenon, which is very common in Austronesian languages, has important 
implications for formal theories that take the lexical categories N and V as 
fundamentally different. It also raises analytical questions, such as how to 
account for the surface homophony between finite forms and nominalised forms 
in (especially) the Eastern Austronesian languages.

In Tukang Besi there is much more verbal morphology than there is nominal 
morphology. In the valency-increasing category there are three causative 
prefixes and three applicative markers (chapters 9 and 10), and 
valency-decrease is accomplished by three passive-like prefixes, two 
reciprocal prefixes, and various other prefixes (chapter 11). Object 
incorporation also features in the language as a valency decreasing device, 
and two varieties of possessor ascension conspire to complicate the picture 
of valency and grammatical relations (chapters 7 and 20). Serial verb 
constructions (chapter 8) are used to express adverbial notions, as well as 
aspect and modality.


This is an excellent grammar, but I do have a couple of comments.
One flaw of the book is that the fieldwork and the database on which it 
is based have not been documented. Obviously, a grammar based on, 
say, 3 months of fieldwork and a 3-hour spoken text corpus is much less 
reliable than one based on 18 months of fieldwork and a corpus of 15 hours. 
>From this grammar, it is not clear how much time the author spend in the 
field 'between 1992 and 1995', where he spent it, with whom he worked, and 
for how long. With respect to his data base, the author states that most 
materials used in the grammar have been taken from recordings of traditional 
stories, process descriptions, and recordings of conversations of various 
speakers, but again the exact data are missing: how many hours of texts were 
recorded and transcribed, and who were the speakers? Another flaw is that it 
is not made explicit that Lexical Functional Grammar is used as the major 
descriptive tool. Especially for readers that are not familiar with LFG it 
would have been helpful if the book contained a brief introductory section, 
explaining the major differences between generative and declarative phrase 
structure (the latter of which is used in this grammar), and the formal 
notations of e.g. argument structure and valency-changing morphology. 

Nevertheless, A grammar of Tukang Besi is an exemplary reference grammar. The 
author has a thorough knowledge of Tukang Besi, and shares this by coupling 
acute analyses of the data with coherent presentations. The illustrative 
material provided in the text is generally well-chosen: apart from clarifying 
the analysis, in many cases it is sufficiently rich to leave the reader 
something to discover on his/her own. I am confident that this grammar will 
prove to be highly valuable in the testing of hypotheses of linguistic theory 
on e.g. argument marking and valency-changing morphology, to the further 
development of descriptive (Austronesian) linguistics, and to linguistic 
typology in general. 

Marian Klamer is a fellow of the Royal Dutch Academy of Arts and Sciences, 
currently affiliated to the Departments of Linguistics of the Vrije 
Universiteit in Amsterdam and Leiden University, in The Netherlands. She has 
published various articles and a grammar of the Eastern Indonesian language 
Kambera (1998, A grammar of Kambera, Mouton de Gruyter), as well as articles 
on the related languages Leti, Buru, and Tukang Besi. Her research interest 
is language contact and change in Eastern Indonesian languages, with 
particular focus on the interface lexicon-morphosyntax.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue