This chapter attempts to do two things. Firstly, to examine lexical and grammatical relationships between the dialects of Dyirbal and surrounding dialects in an attempt to assess the genetic relationships of the dialects and to reconstruct past tribal movement. Secondly, to assess ways in which individual dialects of Dyirbal differ from the general pattern of the language, with a view to internal reconstruction of some aspects of past stages of Dyirbal.

As mentioned in 1.1, it seems likely that most Australian languages (certainly, all those outside Arnhem Land and the Barkly Tableland) are genetically related; we can call this a ‘weak genetic relation’. In this chapter we are interested in whether a pair of languages are ‘strongly genetically related’ – that is, whether they can be related to a common ancestor that is considerably less ancient than proto-Australian. If two languages belong to the same ‘branch’ or ‘subgroup’ of a family then they are said to be ‘strongly related’; if they are members of different branches then they are ‘weakly related’. Consider an Indo-European analogy: Welsh is strongly related to Breton and Cornish, and (slightly less strongly) to Irish, within the Celtic branch of Indo-European; it is only weakly related to languages such as Lithuanian, English and Greek.

In the discussion below we will use ‘genetic relationship’ to mean ‘strong genetic relationship’. Thus we conclude that Dyirbal and Yidin, the language next to the north, are probably not genetically related; that is, to connect them on the ‘family tree’ of Australian languages one would probably have to pass through proto-Australian, at the apex.

Morphology of nouns and adjectives

Stem-forming affixes. There are about a score of affixes in Dyirbal that form nominal stems from nominal roots; some of them also function as affixes to members of other word classes. Some of the affixes are important syntactically – for instance -bila ‘with’ as a type of possessive, and -gara and -maŋgan as coordinators; others merely provide semantic qualification of the noun or adjective they occur with. NPs involving certain of the affixes can make up complete sentences; this is not so in the case of other affixes.

The full functional possibilities of each affix are described below. The section ends with a discussion of the use of bound forms dayi galu, etc. (3.2.3) as nominal affixes.

[1] -ŋa This affix has already been discussed in 3.2.1 and 5.8.3. It is listed as a stem-forming affix, rather than as a case inflection, because [i] it is always optional with nouns; and [ii] it can be followed by other stem-forming affixes, -ɲ;a is generally added only to a nominal referring directly to a particular person or persons – a proper name, or else a noun used in a particular instance as a proper name.

This chapter surveys, rather briefly, some of the recurring linguistic features of languages across the continent. It does not purport to be a complete ‘Handbook of Australian languages’ and does not attempt to list the many individual variations from the basic pattern.

General

There were perhaps 600 aboriginal tribes in Australia at the time of the European invasion at the end of the eighteenth century [Tindale, 1940]. Each tribe spoke a distinct dialect, which usually had considerable lexical and grammatical similarities to neighbouring dialects. The existence of extensive dialect chains makes it difficult to put an exact figure on the number of distinct aboriginal languages; recent guesses have been that there were about 200 [Wurm, 1965; O'Grady et al., 1966].

On the basis of the similarities summarised below it has for some time seemed likely that at least the majority of Australian languages are genetically related.

Chapters 3, 4 and 6 present the basic ‘facts’ of the grammar of Dyirbal. This chapter attempts to interpret, generalise from, and explain the basic facts; the discussion is thus at a higher level of abstraction, and is more speculative and arguable, than chapters 3, 4 and 6.

Points to be explained

In the last chapter we mentioned that certain types of sentence were possible but that other, similar types were impossible; that certain combinations of processes were acceptable while others were unacceptable; and that certain inflections which appeared to be rather different in function were identical in form. In no case did we attempt anything more than a limited and ad hoc explanation of these facts. Before we embark on an investigation of the underlying syntactic nature of Dyirbal, it will be useful to list some of the points from chapter 4 that appear to be in need of explanation:

1. [1] The fact that word order is quite free, except in the case of some particles, and when there is multiple embedding. It should be noted that it is the order of WORDS in sentences that is free, not just the order of phrases (4.5.4, 4.15.3, 7.8).

2. [2] The apparent conflict between the nominative-accusative paradigm of pronouns, and the nominative-ergative syntax of nouns, adjectives, and noun markers (3.2.1, 3.3.1, 4.1).

3. [3] Case inflections on pronouns – in particular, the difference between the inflections on singular pronouns in G, and in D and M (3.3.1).

4. […]

A full account of the semantics of Dyirbal would be considerably more lengthy than the description of the grammar and phonology, in chapters 3–7. In this chapter we refer briefly to some of the main points of Dyirbal semantics, paying attention to aspects that reinforce and help to explain parts of the syntax. A complete account of the semantics, in the form of a comprehensive dictionary-thesaurus of the language, is in active preparation. A provisional statement of verb semantics, going into considerably more detail than is attempted here, will be found in Dixon [1971].

Guwal-Dyalŋuy correspondences

We mentioned in 2.5 that when a speaker was within earshot of a ‘taboo’ relative he had to use Dyalŋuy, the so-called ‘mother-in-law’ style. In all other circumstances he had to use Guwal, the ‘everyday’ style. Every dialect has its own Guwal and Dyalŋuy. Each Dyalŋuy has identical phonology, and almost exactly the same grammar, as its Guwal. However, it has an entirely different vocabulary, there being not a single lexical item common to the Dyalŋuy and Guwal of a tribe.

Confronted with a Guwal word a speaker will give a unique Dyalŋuy ‘equivalent’. And for any Dyalŋuy word he will give one or more corresponding Guwal words. It thus appears that the two vocabularies are in a one-to-many correspondence: each Dyalŋuy word corresponds to one or more Guwal words (and the words so related are in almost all cases not cognate with each other).

What follows is a revision of the sections on grammar and phonology of the writer's London PhD thesis [Dixon, 1968a]. The main points of the semantics section have already been summarised in print [Dixon, 1971]. Further work on the lexicon, and its semantic structure, with particular reference to the special ‘mother-in-law language’ (2.5, 8.1), is proceeding, with a view to the eventual publication of a comprehensive dictionary-thesaurus of Dyirbal.

Since the grammatical natures of Australian languages are not widely known, it has seemed worthwhile, in chapter 1, to give a brief survey of some of the recurring characteristics of languages across the continent. In addition, some references to points of similarity in the grammars of other languages are included, in smaller type, throughout the description of Dyirbal.

The grammar is written at two distinct ‘levels’. The ‘facts’ of the grammar – affixes, their syntactic effect, types of construction, and so on – are described in chapters 3, 4 and 6. Chapter 5 interprets some of these facts, setting up explanatory generalisations and describing the ‘deep’ grammar of Dyirbal in terms of a number of syntactic relations and a number of transformational rules. It has seemed desirable to (at least partially) separate facts from interpretations in the case of a language like Dyirbal that has not previously been described in any way. The correctness of chapters 3, 4 and 6 cannot seriously be in dispute. Chapter 5, however, is far more open to argument.

Linguistic type

Dyirbal is a typical Australian language. It is entirely suffixing, largely agglutinative, and has extraordinarily free word order.

Phonologically it is rather simple, having the smallest number of phonemes of any Australian language. In addition to four stop-nasal series – bilabial, apico-alveolar, dorso-velar and laminal – there is a single lateral, a semi-retroflex continuant, a trill and two semi-vowels. Dyirbal has the usual three-vowel system, length not being significant. Roots and inflected words begin with a consonant and have at least two syllables; intervocalically there are two and three-member consonant clusters. The first syllables of roots, and of most affixes, bear stress.

There are clearly defined classes of (first and second person) pronoun, noun, adjective, verb, adverbal, time qualifier, particle and interjection. There are four noun classes (genders), marked not in the form of a noun but by a ‘noun marker’ that normally accompanies a noun; noun markers also have some of the functions fulfilled by third person pronouns in other languages. A noun marker indicates the class of a noun, agrees with it in case, and also indicates whether the referent of the noun is visible or invisible; here or there; up hill, down hill, up river, down river or across river; and whether a long, short or medium distance up or down. Pronouns distinguish singular, dual and plural in both first and second person; there is no inclusive/exclusive distinction, and no incorporation of pronominal affixes into the verb.

Each Dyirbal sentence has both a grammatical and a phonological description. The phonological description consists of a number of ‘phonological words’: each of these is based on a single ‘word class’ root, the phonological form of the root undergoing phonological addition or alteration corresponding to (optional) stem-forming or derivational processes and (obligatory) inflectional processes.

The word as a phonological unit is marked by an initial stressed syllable, final inflectional ending, and a fixed form: that is, the ‘constituents’ of a word cannot be permuted within the word. There is, however, considerable freedom of the order in which words can appear in a sentence (7.8).

Each word is a sequence of phonological segments, a segment being a set (or ‘bundle’) of phonological features; each feature belongs to a phonological system. A single statement of possible word structures can be given (covering words of all grammatical types). The possible phonic realisations of each phonological feature are described (relative TO the realisations of the other features in the system).

Dyirbal words tend to be phonologically rather compact: that is, certain phonological processes of insertion, assimilation, elision and conflation act to ensure that, say, a consonant cluster is not only one of the ‘possible’ clusters for the language, but is also one of the more frequent or ‘popular’ clusters (7.5). There is no phonological cohesion (sandhi and so on) between words.

As part of what appears to be a continuing campaign to denigrate Australian languages (thereby demonstrating the supposed lack of intellect of their speakers) it has been pointed out that ‘logical connectors’ – ‘or’, ‘because’, ‘if, and so on – are frequently lacking. Love [1938: 119] remarked of Worora, ‘Alternative propositions do not occur. The Worora man does not say, Shall I do this or that? He says, Shall I do this? No. I shall do that. If asked in English, Will you have this or that? he will invariably reply, Yes, desiring both.’

Dyirbal does not have any particles exactly corresponding to English ‘or’, ‘if’ or ‘because’ (any more than English has words corresponding to Dyirbal mugu, biri or wara). The language is, however, quite as capable as English of expressing conditions, alternatives and implications. Amongst the syntactic means employed are:

ENTAILMENT

The major construction here is that involving an implicated VC, in purposive construction – examples were given in 4.4.3, 5.3.3–4. An implicated VC refers to an action (action) that is the ‘consequence’ of the action (action) referred to by the preceding NP. There are two possibilities: (a) action was done deliberately, just so that action should be possible – e.g. run to a tree to climb it, hit someone to kill them; (b) action is the unavoidable (but unplanned) consequence of actio – e.g. perso goes to a place and as a result encounters person who (unknown to person had settled at the place (see text xxv, line 40, p. 391).