Professional Documents
Culture Documents
net/publication/242072708
CITATIONS READS
8 227
1 author:
David Rose
The University of Sydney
55 PUBLICATIONS 2,118 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by David Rose on 30 January 2017.
Pitjantjatjara Processes:
An Australian Experiential Grammar
David Rose
In R. Hasan, D. Butt & C. Cloran (eds.) Functional Descriptions: Language Form & Linguistic
Theory. Amsterdam: Benjamins, 287-322
1. Introduction
This paper interprets one dimension of the grammar of an Australian language, Pitjantjatjara, from
a functional perspective. Pitjantjatjara is the first language of several thousand anangu people in
central Australia, most of whom are also multilingual in dialects of indigenous languages and
English. It is one amongst many dialects of the language classified as `Western Desert', spoken in
an arc of arid lands from the coast of north-western Australia to the Great Australian Bight in South
Australia. The people of this area are related not only by their language, but through a system of kin
and marriage relations that are established and renewed in a network of ceremonial cycles, to
participate in which, vast distances are travelled by the peoples. Between the 1930s and 60s, almost
all the western desert peoples moved into missionary or government run settlements. Since the
1970s they have assumed control of these communities, and established new ones within their
traditional lands. While English literacy has been neglected in the past, there is now a strong push
by the Pitjantjatjara people in its favour in school and adult education.1
1 E.g. "The Pitjantjatjara Yankunytjatjara Education Committee are pushing the importance of English with
our school students. Our committee has a policy which asks all non-Anangu staff in schools and people working in our
communities, to only speak English to the school students. We want this to happen both inside and outside of school"
(PYEC, 1992).
PITJANTJATJARA PROCESSES
2
The description offered in this paper makes comparisons at the levels of language and
meta-language. Pitjantjatjara grammatical functions and their structural realisations are compared
and contrasted with Halliday's 1994 functional description of English; and this functional approach
is compared with the formalist tradition which has been the dominant paradigm for the descriptions
of Australian languages to date. An academic description of indigenous languages is implicitly
comparative in two senses: first the indigenous meanings and forms are translated into English at
the same level of abstraction as wording in examples, and secondly, at higher levels of abstraction a
translation of descriptive categories takes place. In this paper, both kinds of comparisons will be
made deliberately in order to facilitate understanding across languages and linguistic paradigms.
There is considerable potential for enabling teaching of both English and Australian languages if
the practice of teaching is informed by an understanding of how each grammar works in relation to
the other. Bilingual speakers need to know how to express their meanings in either language, and
which meanings may not translate directly. This suggests a descriptive approach that starts with
meanings and asks how they are realised as grammatical structure; in other words, the approach
suggested is a functional one. It also suggests a comparative approach to grammatical function.
Which functions are common to both languages, which differ, and how are they expressed as
structures? Beyond the pedagogic applications opened up by such comparisons, there arise
questions about the nature of the relationship between languages as remote in time and space as
these. Traditional descriptions have focussed on universal formal categories of languages, sounds,
word classes, verb conjugations, nominal cases and so on. From these perspectives Pitjantjatjara
appears to resemble Latin more than English, since they both favour suffixing and traditional
descriptive categories tend to derive from Latin morphology. A systemic functional comparative
approach will establish interlanguage relationships in both meaning and structural realisation
simultaneously, at various degrees of delicacy. This enables us to pinpoint where the languages
converge and where they diverge, and to speculate on possible explanations.
2 See for example, Norman Pama Historical Phonology (Blake 1980), The History of Australian Languages: A First
Approach (Capell 1979), The Languages of Australia (Dixon 1980), Proto-Ngayarda Phonology (O'Grady 1966),
Preliminaries to a Proto Nuclear Pama-Nyungan Stem List (O'Grady 1979), and Classification of Australian
Languages, including Tasmanian (Wurm 1971).
3 For example, Trudinger 1943, Douglas 1964, Glass & Hackett 1970, Institute for Aboriginal Development n.d,
Goddard 1985.
PITJANTJATJARA PROCESSES
4
types of clause structure AOV/SV), with word rank classification (classes of verbs
transitive/intransitive).
As a formal description Dixon's binary classification of verb types is under-elaborated since
clauses with more than one complement (object in traditional terms) are common in Australian
languages, as are also clauses without a verb, and complements may be projected clauses as well as
nominal groups. The "core" participant functions A, O and S are classified on morphological
grounds, inflecting as "ergative", "accusative" and "absolutive case" respectively. However it is
common for two nominal groups within one clause to take the same inflectional form, just as it is
common for clauses to occur with two participants, neither of which is inflected as "ergative".
Dixon's response to these problems is to propose deep/surface structure transformations where the
same meaning may be expressed by two distinct structures, in order to comply with rules of syntax
that are literally 'meaningless'. This dualising of form and meaning reflects the principle underying
formalist accounts, of an arbitrary relation between grammatical structure and semantic concept.
For example a clause with two participants, one inflected as S and the other in "dative" case, is
interpreted as a transformation of a deep AOV structure. Dixon asserts that the "underlying A NP
goes into S function; the deep O NP is now marked by the peripheral syntactic case, dative" (Dixon
1980). Working within Dixon's formalist framework, Goddard (1985) adopts the label
"ambitransitive" to expand the classification to accommodate Pitjantjatjara verbs which function in
both transitive and intransitive clauses, and "ditransitive" for verbs that occur in transitive clauses
with more than two participants, but these remain statements about structure only.4 In the
interpretation proposed in this paper, these regular structural variations are motivated not simply
and primarily by rules of syntax but by differences in meaning.
As a functional description the "SV/AOV" binary system is over-generalised because it is
based on only one type of process. Only material process clauses in Pitjantjatjara typically display
the distinction between `Actor doing' and `Actor doing to/with Goal'. Although material process
verbs in Australian languages tend to function in either transitive or intransitive clauses, there are
many verbs which may function in both types, particularly verbs of sensing and saying, as well as
doing. The intransitive/transitive choice is more accurately a clause rank system. For example the
verb wangkanyi `saying' may function as a verbal process with two or three participants, viz. Sayer,
Receiver and Verbiage, realised as nominal groups. Unlike material processes, verbal ones may
project locutions, while mental processes may project ideas, in which case Sayer may be the only
nominally realised participant. Furthermore the same verb may appear as a behavioural process,
construing verbal behaviour as a material act, with only one participant, a Behaver, as in English
'Mary is talking'. Since relational clauses in Australian languages are often realised without a verb,
these are frequently classed as noun phrases rather than as clauses in formalist descriptions,
particularly since both participants typically take the "accusative" case inflection, and so are not
morphologically distinct. A rich semantic domain in Pitjantjatjara is elided in such accounts,
restricting possibilities for comparison with English relational processes.
These nominal forms distinguish participant and circumstantial functions between multiple nominal
groups in clauses. For example, in intransitive material processes Actor takes the unmarked form.
The unmarked form for common or proper nominals is NEUTRAL, while that for pronouns is
ACTIVE. In transitive material processes, Goal takes NEUTRAL form, while Actor takes ACTIVE
form, irrespective of whether either of these participants is a common/proper nominal or a pronoun.
The difference in inflection patterns between pronouns and other nominals is a product of
their relative frequency in different functional roles. Pronouns most frequently function as the
participant that 'actualises' the process, Actor, Senser, Sayer, Carrier, Token. Since Pitjantjatjara is
5 See further discussion of case and function in the context of Tagalog, in Martin, this volume.
PITJANTJATJARA PROCESSES
6
a purely spoken language, pronouns are the most common realisation of these roles, referring to the
interactants (you, I & we) or other people in the contexts of situation or preceding discourse (s/he,
they). Because these are their typical roles, the unmarked inflection for pronouns is ACTIVE.
Common nominals on the other hand, which denote classes of entities, equally likely to occur in
'actualising' roles, as in any other. In clauses of two or more participants, they are more likely to
occur in 'non-actualising' roles. Common nominals therefore take NEUTRAL case as their unmarked
form. Proper nominals, naming people and places, follow the same ACTIVE/NEUTRAL pattern as
common nominals, except that their NEUTRAL case is overtly inflected. This is because their
uninflected form functions interpersonally as Vocative.
Duration iterated imperfective process ankula ankula ankula 'going going going'
Location:rest nom gp-ngka, -la, 'at, in'
motion 'towards': nom gp-kutu; 'away from': -nguru; 'through': -wanu
Means nom gp-gnka, -la 'with (thing)'
Destination nom gp-ku 'for (place)'
Intention nom gp-kitja 'for the purpose of (thing)'
Reason nom gp -nguru 'from (thing)'
Behalf nom gp -ku 'for (person)'
Accompani. nom gp-ngka, -la 'with (person/thing)'
Addition nom gpˆkulukulu 'as well as (person/thing)'
Quality adverbial gp
Comparison nom gp-purunpa 'like (person/thing)'
Role nom gp 'as (role)'
The circumstantial functions of causative and locative suffixes are similar to those of prepositions
in English,6 and they cover a remarkably similar semantic spread. The major exception to this in
Table 2, is the realisation of Duration as a series of imperfective non-finite processes.7 These may
consist of up to six iterated processes, the greater the number, the longer the Duration. These series
are not part of the verbal group in the clause, but are distinct elements, frequently spoken on a
separate tone group, presenting them as marked New information, on the same pattern as nominal
6 Describing a similar system in Kuniyanti, another Australian language, MacGregor (1992) labels such suffixes as
'post-positions', drawing attention to their functional similarity with prepositions in English.
7 'Serial verb form' in formalist accounts.
PITJANTJATJARA PROCESSES
7
circumstances. This construal of Duration is independent of any temporal divisions, which are the
basis of the English construal of Duration, exact or inexact. It strongly recalls Whorf's description
of temporal representations in Native American languages, as 'subjective experience of time
becoming later', in contrast to the 'objectification' of time in European languages.
2. Material processes
These examples illustrate the construal of material action in Pitjantjara, as either effective
(transitive), as in (3), or non-effective (intransitive), as in (1) and (2). Most Pitjantjatjara verbs only
function in one or the other type of clause, but there are still a good number of verbs which may
function in either, as exemplified in (4):
8 The "case" and other grammatical marking for the constituents in the Pitjantjatjara clauses is shown in capital
letters along with the literal translation. The key to the terms is provided at the end of this paper.
PITJANTJATJARA PROCESSES
8
The same verb that functioned effectively in (3) involves here in (4) only an Actor; the
process is not extended to a Goal. The `shower' is not a participant but a circumstance associated
with but not inherent in the process. Note however that the Actor remains the participant that
initiates the action; it is not possible for the Goal of the effective process to become the Actor in the
corresponding non-effective process, in contrast to the English pattern, as exemplified by the pair
`she is washing the clothes' and `the clothes are washing'. The apparently equivalent Pitjantjatjara
clause ultja paltjini, literally `clothes washing', actually involves an implicit (ellipsed) Actor `(s/he)
is washing the clothes'.
This is a significant point since it illustrates the model of nuclear transitivity in
Pitjantjatjara. In the environment of material processes, the nuclear configuration is Actor/Process
(+Goal) in contrast with English, where the nuclear model is Process/Medium (+Agent). Whereas
the English model focuses on the causation of the process, the Pitjantjatjara model focuses on its
resulting effect.
In this example, both nyantju `horse' and puli `hill' take the neutral form. Their respective functions
are semantically distinct without any need for structural distinction, i.e. the `horse' is clearly the
Actor in the process of 'climbing'. However, the `hill' is not a circumstance of location; had it been
a locative circumstance, it would have displayed a locative suffix (cf example 7 below). So
although the process is extended to a second participant, the Actor is not marked in active form.
What then is the nature of the second participant and how does its relationship to the process differ
from that of a Goal? The same verb tatini `climb' can function in a transitive clause in which the
Actor is in fact inflected in active form:
These alternative structures are not random but realise a semantic distinction. In example
(6), the more common structure for effective material clauses, the process is not only extended to
but also results in an effect on the Goal. By contrast, in example (5), the process is extended to the
PITJANTJATJARA PROCESSES
9
second participant, but does not result in an effect on it. This suggests that where common/proper
nominals, as Actor in effective material clauses, are not inflected in active form, this marks the
process as `non-effective'. That is to say, the process in such clauses does not produce an effect on
the second participant. This participant, to which the process is extended without a resultant effect,
is labelled Range (following Halliday 1994:146-9). Whereas in English, Range is distinguished
from Goal by its potential for realisation as a prepositional phrase, in Pitjantjatjara a similar
function is distinguished by the lack of active inflection on the Actor.
These examples are further evidence of the inadequacy of the formalist definition of
transitivity as a binary verb classification. Despite its common occurrence in transitive clauses,
Goddard (1985) classifies tatini `climb' as an "intransitive verb", giving the following example
(analysis ours):
This clause is translated by Goddard as "Those kids over there are climbing a tree", implicitly
interpreting the clause as an Actor/Process/Range configuration. But in this example `tree' is not
inflected as Range, but as Circumstance of location. The experiential focus is the Actor + Process
`children+climbing' nucleus; its domain (`tree') is more peripheral than if it had been marked as
Range.
There are also a few verbs such as patani `wait', and ngurini `search', which function in
Actor/Process + Range structures, in which the Range carries an additional sense of `purpose'. In
this case the Range is inflected with the `causative' suffix -ku or -mpa ("dative case" in traditional
terms).
Dixon (1980) explains this type of structure as a surface intransitive structure SV+dative,
derived from deep transitive structure AOV, and Goddard (1985) accordingly classifies verbs such
as patani as "intransitive". However there are verbs which occur in both this type of structure and in
effective structures, with Actor-ACTV and Goal-NEUT. These include ngurini, which Goddard
classifies as a "transitive" verb, and panykani `stalking', which he classifies as "ambitransitive"
without suggesting a semantic distinction between the structural variants. However, as we saw in
examples (5) and (6), there certainly is a semantic distinction expressed by these alternative
structures.
PITJANTJATJARA PROCESSES
10
A third participant role in dispositive material processes, such as unganyi `giving' is the
Recipient:
In this case, both Goal and Recipient are realised in neutral form, but their roles are semantically
distinct; it is unlikely that grandma gave us to the food. However Actor is distinguished from the
other participants by the active suffix. Goddard (1985:168) recognises this configuration as
regularly occurring with the verb unganyi, "The thing given and the recipient both take accusative
case", though which he would classify as O in Dixon's system is not clear. Goddard goes on to add
"with the recipient usually coming first, if both are mentioned", and contradicts this with his first
example (analysis ours):
Goddard may be suggesting `word order' to distinguish the two participant functions, but as we
discussed above, this is unnecessary since the two participants are semantically distinguishable
without structural marking. The order of constituents in Pitjantjatjara clauses is textually
significant, not experientially.
Each of the participant roles, Actor, Goal, Range and Recipient correspond closely between
Pitjantjatjara and English. The labels I have used here for participant roles follow Halliday (1994).
In English Actor is distinguished from Goal either by sequence or as a by-Adjunct. The other
participant roles Range and Recipient are also distinguished either by sequence or as prepositional
phrases. Pitjantjatjara relies on affixes and semantic distinctions to achieve the same ends.
a: The inherent participant, the Senser, is a conscious animal or person, whereas participants in material
processes may be any kind of entity.
b: There must be a complement to the process, that which is sensed, which may be represented as a clause rank
participant, a Phenomenon, or as a projected clause, an idea.
PITJANTJATJARA PROCESSES
11
Thus the crucial recognition features of mental clauses include the type of entity that can function
as Senser, and the potential for projection. The model of consciousness construed in mental
processes in Pitjantjatjara includes two general domains, perception of, or reaction to, external or
internal phenomena. In both types, the perceived phenomena may be represented as clause rank
participant, a Phenomenon, or as projected clause(s).
3.1 Perception
The grammar of perception construes the Senser's apprehension of external events, entities and
qualities as identical to her apprehension of phenomena internal to her consciousness. There is no
structural distinction between cognition and perception as we find in English for example. This
conflation of external and internal perception is mirrored lexically in the general mental verb
kuli-ni, which may be interpreted in English as `hear', `heed', `consider', `know', `understand',
`remember' or `feel', depending on the context. In other words the verb kulini may function as a
process of perception of external or internal phenomena, or simply as perceptive behaviour. The
alternative expressions of perception, as a projecting clause complex, and as a simple clause are
presented as follows.
3.1.1 Projecting
Projected perception represents meantal processing as 'internal speech'; the relation between the
projecting mental process and the projected clause is paratactic, the latter represented as quoted
'direct speech'. This is illustrated below in (11) and (12) for internal and external perception,
respectively.
(11) INTERNAL:
1 nganana kuli-nu
we-ACTV did think
Senser Pro:cog
(12) EXTERNAL:
α munu paluru nya-ngu
and she-ACT did see
Senser Pro:perception
The representation of internal perception as direct speech, as in (11), is not dissimilar to to the less
frequent variant of projected cognition in English, as a paratactic clause complex (Halliday
1994:252). But the corresponding representation of external perception is unique; processes of
seeing and feeling phenomena external to the Senser are construed in the same manner as processes
of thinking (about) phenomena internal to her consciousness. That is the general concept of
perception is construed as the Senser 'thinking' in language. This is not the same as 'saying' as it is
construed in verbal processes (see below), which construes speech as an act of social exchange
between Sayer and Receiver. Rather it represents the perceptive domain of consciousness as the
production of meaning in the form of quoted wording. It is virtually impossible to re-present this in
English without reference to verbal processes such as 'talking to oneself', where the Sayer has the
same identity as Receiver, or to spatial metaphors like 'say it in your head'.
(13) PERFECTIVE:
α Jimmy-nya kuli-ni
Jimmy-NEUT is thinking
Senser Pro:cog
‘β anku-ntjikitja
to go
Pro:mat
Jimmy thinks he'll go
3.1.2 Non-projecting
Clauses with mental process are not constrained to project: i.e. they may be non-projecting. In that
case the Phenomenon is realised as a nominal group. It represents an entity or name for a process
that the Senser perceives, either in her external environment or internal thoughts. To distinguish
their roles, the Senser takes the ACTIVE nominal form, and the Phenomenon the NEUTRAL form,
reflecting the pattern in effective material clauses.
3.2 Perception
PITJANTJATJARA PROCESSES
13
Mental processes of perception are like cognitive processes in that, here too, the Senser is a
conscious being, and that which is perceived may be a thing or projected idea. Projection in
perception differs from cognition in two ways:
a: Projected perceptions are only hypotactic, i.e. they do not represent perceptions as wording (cf example 11i-ii
above).
b: Projections are imperfective only, i.e. realis, since it is not possible to perceive events occurring in the future.
(15) NON-PROJECTING:
tjilpi-ngku malu nya-ngu
old man-ACTV kangaroo-NEUT did see
Senser Phenomenon Pro:perception
the old man saw a kangaroo
(16) PROJECTING:
α nganana nya-ngu
we-ACTV did see
Senser Pro:perception
‘β nyuntu wati-pitja-nyangka
you-ACTV across-go-IMPERF:SW-S
Actor Pro:mat
we saw you going across
(17)
α ngayulu kuli-nu
I-ACTV did hear
Senser Pro:perception
To this point the semantic potential of mental processes and their similarities and
differences to material processes are comparable to those for English as described by Halliday
(1994). For both languages, material and mental processes differ with respect to number of inherent
PITJANTJATJARA PROCESSES
14
participants, the consciousness of participants, and the type of complementation possible, i.e. as a
thing or as a projected idea.
However processes of doing and perceiving also resemble each other in another respect: in
both cases, there is potential for choosing either one, two or more participants realised as nominal
groups. Where there is only one inherent nominal participant, i.e. Actor in material or Senser in
mental, it will take the unmarked form. By contrast, where there are two, i.e. Goal in material or
Phenomenon in mental, these will take the unmarked NEUTRAL nominal form, while Actor or
Senser is distinguished by the ACTIVE form. Despite these similarities, it must be emphasised that
the Actor/Process+Goal material configuration is semantically distinct from the
Senser/Process+Phenomenon mental figure. The material model is clearly one of effect; an
effective material process results in some kind of effect on the Goal, creating it, changing it in some
way, transporting or exchanging it. But the mental model is one of signification; perceptive
processes re-present a concrete entity or process as an abstraction, an idea internal to the Senser.
This is made explicit in projections, which represent perception as internal speech, but applies
equally to Phenomenon:things. The perception has no material effect on the concrete thing
represented by Phenomenon; rather it results in a semiotic entity that is abstracted from the material
entity it signifies.
In formalist descriptions, the differences between material and mental process types are not
recognised. The distinct functions Goal and Phenomenon are both labelled O, since they are
realised by a common nominal form, so-called "accusative case", while Actor or Senser are both
labelled A, with "ergative case" inflection. However, insofar as "ergative case" implies a single
functional interpretation, as Agent of the process, it is over-generalised. We have already
demonstrated that the material model in Pitjantjatjara is one of effect, rather than agency, and that
effectivity is a clause rank system distinguishing non-effective Actor/Process configurations from
Actor/Process+Goal effective ones. In the latter, the so-called ergative inflection on Actor is more
accurately an ACTIVE form, which functions to distinguish Actor from Goal in the structure as a
whole; it does not inherently ascribe agency to the entity functioning as Actor. This is more
apparent in mental processes, in which the Senser is clearly not agentive; her process of perception
does not cause any change in the perceived entity or process. Rather, like Actor, the Senser is the
participant which 'actualises' the process, and is consequently the one inflected in ACTIVE form.
3.3 Reaction
Mental reaction differs from perceptive processes in its construal of the nature of the process and its
relation to its Phenomenon. We have seen how perceiving is construed as extended from the Senser
+ Process configuration to the Phenomenon, in which they are somewhat similar to transitive
material processes. Reaction on the other hand is construed as a change in the Senser in response to
an outside stimulus. This distinction recalls Halliday's description of bi-directionality in English
mental processes, `I like it/it pleases me' (Halliday 1994), and Martin's formulation for Tagalog
processes, of a "centrifugal" vs "centripetal" orientation (no date).
PITJANTJATJARA PROCESSES
15
Reaction processes in Pitjantjatjara are realised by verbs of affective or cognitive states that
are inflected for inceptive aspect, suggesting an internal event of `becoming'. The Phenomenon here
is realised by the CAUSATIVE nominal inflection (which also inflects circumstances Location:
towards or Destination), or a perfective non-finite process, e.g. anku-ntjikitja 'to do' (the suffix of
which also inflects circumstances of Intention). These nominal and verbal forms both suggest an
external stimulus which the internal event is directed towards. Reaction projections are, in other
words, always hypotactic and non-finite:perfective. A relatively large group of Pitjantjatjara verbs
function in reaction processes. The majority of these represent affective states. Below are some
instances of these, followed by examples of non-projecting and projecting reaction:
(18) NON-PROJECTING:
i nyuntu nyaa-ku ngayu-ku pikari-nganyi?
you-ACTV why? me-CAUSTV being angry
Senser Phenomenon Pro:reaction
why are you angry at me?
(19) PROJECTING:
α anangu tjuta mukuringku-pai
person many-NEUT does like
Senser Pro:reaction
‘β Tjulu-ku anku-ntjikitja
Tjulu-CAUSTV to go
Circ:destin Pros:mat
PITJANTJATJARA PROCESSES
16
many people like to go to Tjulu
In addition to these affective states, there are a few verbs that represent cognitive rather than
affective states, e.g.
nintiringanyi learn
rukuringanyi be reminded
However the structural realisations of these processes are identical to the affective states described
above, so they are construed grammatically as the same process type, despite their lexical
association with cognition. Examples are given below:
(20) NON-PROJECTING:
tjitji tjuta Englisha-ku nintirinku-pai school-angka
child many-NEUT English-CAUSTV do learn school-LOC
Senser Phenomenon Pro:reaction Circ:loc
children learn English at school
(21) PROJECTING:
α ngayulu iriti nintiri-ngu
I-ACTV long ago did learn
Senser Circ:loc:time Process:reaction
‘β motorcar palya-ntjikitja
motorcar-NEUT fix-PERF:SA-S
Goal Pro:mat
I learnt to fix motorcars a long time ago
Many verbs that realise mental reaction also have corresponding causative forms. These
verbs function in clauses which involve an additional agentive participant, the Inducer, as well as a
Senser and Phenomenon; and again the processes may be non-projecting or projecting:
(22) NON-PROJECTING:
tjilpi-ngku ngayu-nya tjukur-ku ninti-nu
old man-ACTV me-NEUT Dreaming-CAUSTV did teach
Inducer Senser Phenomenon Pro:reaction:causative
old men taught me that Dreaming
(23) PROJECTING:
PITJANTJATJARA PROCESSES
17
α kutjupa-ngku palu-nya muku-muku-nu
another-ACTV him/her-NEUT did cause to want
Inducer Senser Pro:reaction:causative
‘β anku-ntjaku
to go
Pro:mat
another person persuaded him/her to go
That which is said may have the status of a participant, the Verbiage, which may be realised
nominally; or it may be a projected clause. Verbal projections may be paratactic, i.e. quoting
speech, or hypotactic, reporting speech. Cross-classifying these two dependency types is the speech
function of the projecting and projected clauses, whether proposition or proposal. Projected
propositions are typically represented as quoted speech, as in mental perception; but they may also
report speech, in which case person and temporal reference concords with that in the projecting
clause, as in English, 'he said that he was coming'. Projected proposals are realised as non-finite
perfective processes, as in mental reaction, 'tell him to go', except where the projecting clause is in
imperative mood, in which case the structure is paratactic and the projected clause is also
imperative in mood, 'tell him give me'. We will begin our analysis with non-projecting clauses, and
then illustrate the paratactic and hypotactic variants of projected proposals and propositions.
4.1 Non-projecting
As with Senser in mental perception, the role of Sayer in non-projecting verbal processes is
distinguished by the ACTIVE nominal form, from the Verbiage which is inflected as NEUTRAL.
There is also a third participant role specific to verbal processes, the Receiver, to whom the saying
is addressed. Receiver is realised by the LOCATIVE suffix, distinguishing its role from Sayer and
Verbiage. Less frequently, if there is no other participant in the clause aside from Sayer, the
Receiver may take the NEUTRAL form. (24) exemplifies some of these roles:
(25)
1 ka ngayulu palu-la wangka-ngu
and I-ACTV him-LOC did say
Sayer Receiver Pro:verbal
(26)
α munu palu-nya tjakaltju-nu
and (she) her-NEUT did say
Receiver Pro:verbal
Note that the projected clause in (26) is introduced with the anaphoric reference item panya, here
functioning as a conjunction which marks the clause complex relation as projection. This is a
general feature of reported propositions, equivalent in this function to 'that' in English reported
speech. Turning to proposals, these may be projected hypotactically as non-finite perfective
processes, or paratactically as finite clauses in imperative mood, where the projecting clause is also
imperative. These are illustrated in (27) and (28) respectively.
(27)
α paluru tjana ngayu-la watja-nu
they-ACTV I-LOC did command
Sayer Receiver Pro:verbal
"β ma-pitja-ntjaku
away-go-PERF:SW-S
Process:mat
they told me to go away
PITJANTJATJARA PROCESSES
19
(28)
1 palu-la wangka
him-LOC tell!
Receiver Pro:verbal
The projected clause in (28) is introduced by the additive conjunction ka. In this environment it
fulfils the same conjunctive role as panya in (26), but here ka is used to mark the clause complex
relation as paratactic projection.
5. Behavioural processes
Behavioural processes resemble mental and verbal processes, in that one participant is typically a
conscious being. However, behavioural processes do not project, and there is most frequently no
other participant, whereas mental and verbal processes must include a complement as participant or
projection. Many verbs that function as mental or verbal processes may also function as
behavioural processes, but the latter represent the processes of sensing and saying as a material act,
rather than the production of meaning:
In the examples, the behavioural process is realised by the verbs `listening', `looking', `talking', etc.,
representing it congruently as an action of the Behaver. But the process of behaving may also be
realised nominally as a second participant, the Behaviour. This is typically the name of a process:
In this example, the meaning of `singing' is realised as a nominal element inma `song'. The process
inkanyi `playing' is required to fulfil the need for a verb to indicate mood, tense and aspect, as in
English `sing a song, play a tune, have a shower' etc. There are a few such verbs in Pitjantjatjara,
with a specific range of Behaviours associated with them, such as arkani `acting, behaving'. The
same structure occurs with verbs of speaking as Process and language varieties as Behaviour:
Both Behaver and the Behaviour take the same neutral form here; since they are semantically
distinct, there is no need for a structural distinction. In this respect, behavioural processes are like
middle material processes with Actor and Range. With the latter, however, the Actor may be any
kind of entity and the Range does not represent a behaviour of a conscious or talking entity.
6.1 Identifying
In Pitjantjatjara identifying clauses, the two inherent participant roles conform with Halliday's
formulation for English (1994:124): "one element will be the Value (meaning, referent, function,
status, role) and the other element will be the Token (sign, name, form, holder, occupant)". The
Value is always a nominal constituent, as is the Token, except in the case of possessive identifying
relations, when it may be inflected as an inceptive process. Identifying relations are illustrated
below for intensive, possessive and circumstantial variants.
6.1.1 Intensive
The intensive identifying relation is concerned with identifying an entity with some role, status,
meaning etc.
There is no structural distinction between Token and Value, unlike English where the Token may
be realised as a by-Adjunct, as in 'Hamlet was played by Mr Garrick' (Halliday 1994:123). But the
two participants are semantically distinct, since Token is the signifier of the Value. In (33) a name
identifies a person; in (34) a person is equated with his social role, and in (35) a place is equated
with its role in the Pitjantjatjara mythological system `place-of-fire'. Identifying relations may be
probed with ...nyaa `which?' and ngananya `who?', e.g.:
PITJANTJATJARA PROCESSES
22
Intensive identifying relations may also be brought about by an agentive participant, the
Assigner. In this case the identifying relation is realised as process, the Assigner takes ACTIVE
form, while both Token and Value are NEUTRAL:
Some verbs functioning in such causative identifying processes, such as inini `naming' or
wangkanyi `saying' are variously classified by formalists as "transitive" or "ditransitive" to account
for the presence of two participants in neutral form.
6.1.2 Possessive
The possessive identifying relation is concerned with identifying a possession with its possessor. In
this type, the Token takes the possessive -ku/mpa suffix:
The possessive type may also be construed as a process, for the purpose of expressing inceptive
phase, together with tense or aspect, by an inceptive suffix on the Token, as in (40):
6.2 Attributive
Attributive relations ascribe an Attribute, a class, possession or location, to an entity, functioning as
Carrier. As with identifying processes, attributive relations may be realised without a verb by
juxtaposition of two participants functioning as Carrier and Attribute. However, they are distinct
from identifying relations in that an Attribute represents a class or descriptive category to which the
Carrier is attributed, whereas the identifying relation equates rather than classifies.
(44) POSSESSIVE:
ngayulu nyantju tjuta-tjara
I-ACTV horse many-POSS
PITJANTJATJARA PROCESSES
24
Carrier Attribute:possession
I have many horses
(45) CIRCUMSTANTIAL:
paluru ngura nyara-ngka
s/he-ACTV place yonder-LOC
Carrier Attribute:location
s/he is at yonder place
(46) INTENSIVE:
kuka nyangatja una-ri-ngu
meat this-NEUT has become bad
Carrier Attribute:process
this meat has gone bad
(47) POSSESSIVE:
walawuru rabbita kanyi-ni
eagle-NEUT rabbit-NEUT is holding
Carrier Attribute Pro:attrib:possessive
the eagle has a rabbit
(48) CIRCUMSTANTIAL:
minyma panyatja wali-ngka nyina-ngi
woman that-NEUT house-LOC was sitting
Carrier Attribute:loc Pro:attrib:circumstantial
that woman was in the house
Like identifying intensive relations, the attributive intensive relation may also be causative
with an additional agentive participant, the Attributor. In this case the Attribute is inflected as a
process, Attributor is inflected in ACTIVE and the Carrier in NEUTRAL form. These are of two
types; the first is related to material processes in that the Carrier is classified by a change in form;
PITJANTJATJARA PROCESSES
25
the second is related to verbal processes in that the Carrier classified by a change in vocation. These
are exemplified in (50) and (51) respectively:
[FIGURE 1 HERE]
The view presented in the Figure illustrates the following topological relationships:
a: Non-projecting verbal processes and mental processes of perception resemble transitive material processes. In
these cases the Senser or Sayer resembles an Actor, since both are inflected in ACTIVE form, although
semantically Senser must be conscious, and Sayer human, while Actor may be any entity. Likewise
Phenomenon or Verbiage morphologically resemble a Goal, although semantically they stands for an
abstraction, a perception or speech act, while Goal represents the concrete entity an action has an effect on.
b: Mental processes of reaction resemble attributive relations where the Attribute is a reactive mental state
(knowledgeable, fearful, etc.), since both Senser and Carrier are conscious things, and both may project.
However non-projecting reaction clauses must involve a Phenomenon which the Senser reacts to, while the
corresponding attributive relations typically don't.
c: Relational clauses in which the relation is expressed as a process resemble material processes. In
circumstantial and possessive relations, the process is realised by verbs of 'posture' or `holding' which can
otherwise function as material processes. Causative attributive and identifying clauses structurally resemble
efective material clauses, since the Assigner or Attributor is inflected in ACTIVE form like an Actor, while the
Carrier or Value resembles a Goal with NEUTRAL inflection. However causative identifying clauses include a
third participant, the Token, and semantically the process is one of assigning an identity to the Token.
Likewise the meaning of causative attributive processes is to attribute a quality or class to a Carrier, and the
Attribute is inflected as a process.
10 See Hasan and Fries (1995) for some discussion of grammatical comparison across languages.
11 cf Ruhlen date, on language typology research using lexical evidence.
PITJANTJATJARA PROCESSES
28
should share, for example, many of its circumstantial representations of space and time with
Pitjantjatjara, whose speakers had maintained their nomadic harvesting economy until a mere half-
century ago. On the other hand, the general semantic domains of material, mental, verbal and
relational processes are arguably an analysis of experience generated by the nature of our
experience of ourselves interacting with our social and material environment. They are likely to be
found in some configuration across cultures.
This should not be read as an attempt to posit language universals in the traditional sense.
This has been pursued and adequately proved in the realms of phonology and morphology by
formally oriented Australianists. What we need to focus on are relationships of similarity and
difference between these languages which now share a single national and local community. And
there are significant differences. English for example has similar structural realisations across
mental processes; cognition, perception and affection differ most markedly in the type of ideas they
project, propositions vs proposals (Halliday 1994:257-60). In Pitjantjatjara, cognition and
perception are conflated, and have quite distinct structural realisations from mental reaction;
perception and reaction are more transparently alternate models of consciousness.
More abstractly, the evolution in English of the ergative model of nuclear transitivity
(Martin 1992, Halliday 1994, Mathiessen in press) contrasts with the nuclear model in
Pitjantjatjara. In the English model the clause nucleus consists of Process+Medium, where
Medium is the inherent participant "through which the process is actualised" (Halliday 1994:163).
English Medium may conflate with alternate transitive roles, such as Actor in middle clauses' the
clothes are washing', or Goal in effective material clauses, 'the clothes are being washed'. In the
latter Actor is optional, and conflated with the nuclear role of Agent, 'by the woman'. In the
Pitjantjatjara nuclear model, the Medium is the inherent participant that 'actualises' the process or
relation, Actor, Senser, Sayer, Carrier, Value. The clause nucleus consists of Actor/Process,
Senser/Process, Sayer/Process, Carrier/Attribute or Value/Token (Attribute or Token may also be
inflected as processes), and it is these 'actualising' roles that take unmarked ACTIVE form as
pronouns. In material clauses the Actor/Process configuration may result in an effect on the Goal,
depending on the effectivity of the process. The Pitjantjatjara material model foregrounds effect
over causation, and this tendency is also reflected in the fact that cause:reason is a minor semantic
motif in the grammar and discourse, whereas cause:purpose is diversified and frequently deployed
(Rose 1993). This contrasts with recent developments in English, which have produced a
proliferation of resources for expressing reason, including the conflation of cause with agency in
circumstantial identifying processes, on the model of 'Value y is caused by Token x' (Halliday
1994:398-403).
The most radical contextual and semantic differences between English and Pitjantjatjara
have arisen over the past five centuries, as English culture has undergone mercantile, imperial and
industrial revolutions. English maintains the core transitivity system that served the needs of its
agricultural/pastoral speakers, but has since developed vast new specialised vocabularies and
potentials for semantic change through grammatical metaphor. A central feature of this semiotic
process is the capacity to nominalise processes which may then function as participants in other
PITJANTJATJARA PROCESSES
29
clauses. Another is to render many types of logico-semantic relation as relational processes, with
myriad shades of meaning. Pitjantjatjara shares neither of these features, which have evolved
primarily in English's written mode, borrowed from Latin and Greek, in the contexts of
mercantilism, administration and science. These processes of grammatical metaphor are major
stumbling blocks for speakers of Australian languages learning to crack the English written code.12
Generally it is less problematic learning to translate spoken Pitjantjatjara experiential meanings into
spoken English grammar, which is predictable given the comparisons we have drawn between
transitivity systems. The experiential semantic potential of spoken English and Pitjantjatjara are
probably broadly convergent; it is when this basic system is played around with in written English
that the semantic potential diverges most dramatically.
Many 19th and early 20th century linguists believed they could make scientifically valid
comparisons between cultures on the basis of the phonology and morphology of their languages.
These methods have been long discredited, not only for their implausibility but because they were
associated with unacceptable theories of ethnic hierarchies. More recently formally oriented
linguists have gone to great lengths to show that all languages, no matter what the technology of
their cultures, share equal levels of phonological and morphological complexity, or that the number
of lexical items used by average speakers of Australian or European languages is about the same.
At the same time there is a fascination with differences in `world-view' between these cultures and
attempts to locate its expression at the levels of lexis or syntax. Typically the latter begin with
philosophical concepts such as `synthetic' (Australian) or 'atomistic' (European) for which evidence
is sought in clause rank structures.13 What these methods lack however is a clear theory of the
complex relation between social context, meaning and grammar, such as has been developed in
SFL over the past 30 years. Whorf's (1938; 1950) elaborate comparisons of the `cryptogrammars' of
native American and `Standard Average European' languages, showed that there were important
differences in their construals of time, for example. Under the ensuing hegemony of universalism
however, his discoveries and methods were rejected or ignored, and it is only recently that the
metalinguistic tools for such large scale semantic comparisons have become available. The most
important recent grammatical work in this respect is not across languages, but between registers of
spoken and written English, many of which are unintelligible to the majority of English speakers
(Halliday 1985, Rose et al 1992, Halliday and Martin 1993).
It is now possible to strike a balance between the impulse towards linguistic universalism
and the interest in cultural difference that reflect complementary movements within modernism.
There is no longer any need to prove or disprove equality or sameness, nor to explain or exoticise
difference. In the emerging post-colonial environment it is necessary to study and understand
cultural and linguistic relationships from simultaneous perspectives of similarity and difference, not
merely to expand the cultural capital of the imperial academy, but to use it for the mutual benefit of
partners in reconciliation.
12 Cf Martin 1991, for a discussion of problems faced by Aboriginal ESL learners, in dealing with grammatical
metaphor.
13 Eg Christie M, 1992. Grounded & ex-centric knowledges: exploring Aboriginal alternatives to western thinking
PITJANTJATJARA PROCESSES
30
Abbreviations
ACTV ACTIVE
CAUSTV CAUSATIVE
circ Circumstance
cog cognition process
destin destination
DURTV durative
FUT future tense
HABIT habitual
INCEPTV inceptive
IMP imperative mood
IMPERF imperfective
intrans non-effective
LOC LOCATIVE
PERF perfective
POSS possessive
PRES present tense
pro Process
PST past tense
react reaction process
Recipt Recipient
SA-S Same Subject
SW-S Switch Subject
TRANS transitive
References
Blake, B. 1980. Norman Pama Historical Phonology. Papers in Australian Linguistics 13, Issue edited by R Rigsby and
P Sutton.
Capell, A. 1979. The History of Australian Languages: A First Approach. Australian Linguistic Studies, Pacific
Linguistics A54, edited by S A Wurm.
Eg Christie M, 1992. Grounded & ex-centric knowledges: exploring Aboriginal alternatives to western thinking. Paper
delivered to the Conference on Thinking, Townsville: James Cook University.
Davidse, Kristin. 1991. Categories of Experiential Grammar. Unpublished doctoral dissertation. Departement
Linguistiek. Leuven: Katholieke Universiteit.
PITJANTJATJARA PROCESSES
31
Davidse, Kristin. 1992. Transitivity/ergativity: The Janus-headed grammar of actions and events. Advances in Systemic
Linguistics: Recent Theory and Practice, edited by Martin Davies and Louise Ravelli. London: Pinter.
Dixon, R.M.W. 1980. The Languages of Australia. Cambridge: Cambridge University Press.
Douglas, W.H. 1964. An Introduction to the Western Desert Language. Oceania Linguistic Monographs 4. Sydney:
University of Sydney.
Fawcett, R.P. 1980.
Fawcett, R.P. 1987. The semantics of clause and verb for relational processes in English. New Developments in
Systemic Linguistics Vol 1: Theory and Description, edited by M A K Halliday and Robin P Fawcett. London:
Pinter.
Glass, A. and D. Hackett. 1970. Pitjantjatjara Grammar: A Tagmemic View of Ngaanyatjara (Warburton Ranges)
Dialect. Canberra: AIAS.
Goddard, C. 1985. A Grammar of Yankunytjatjata. Alice Springs: Institute for Aboriginal Development.
Halliday, M.A.K. 1994. An Introduction to Functional Grammar. London: Arnold.
Halliday, M.A.K. 1985b. Spoken and Written Language. Geelong, Vic: Deakin University Press.
Halliday, M.A.K. and J.R. Martin. 1993. Writing Science: literacy and discursive power. London: Falmer.
Hasan, Ruqaiya and Peter H. Fries. 1995. Reflections on Subject and Theme: an introduction. On Subject and Theme:
From a Discourse Functional Perspective. Amsterdam: John Benjamins
Haraway, D.J. 1989. Primate Visions. London:Routledge.
Institute for Aboriginal Development, n.d. Pitjantjatjara Grammatical Notes. Alice Springs: IAD.
Lemke, Jay. no date. The topology of genre: text structure and text types. mimeo.
Martin, J.R. 1991. Language and Control: fighting with words. Language: Maintenance, Power and Education in
Australian Aboriginal Contexts, edited by C Walton and W Eggington. Darwin: NTU Press.
Martin J.R. 1992. English Text. Amsterdam: Benjamins
Martin J.R. no date. Transitivity in Tagalog: a functional interpretation of case.
Martin, J.R. and Christian Matthiessen. 1991. Systemic typology and topology. Literacy in Social Processes: Papers
from the Inaugural Systemic Linguistics Conference, edited by F Christie. Darwin: Centre for Studies in
Language Education, NTU.
Matthiessen, Christian. in press. Lexicogrammatical Cartography. Tokyo: International Language Sciences Publisher.
O'Grady, G. N. 1966. Proto-Ngayarda Phonology. Oceanic Linguistics 5: 71-130.
O'Grady, G. N. 1979. Preliminaries to a Proto Nuclear Pama-Nyungan Stem List. Australian Linguistic Studies. Pacific
Linguistics A54, edited by S. A. Wurm.
Rose D., 1993. On becoming: the grammar of causality in Pitjantjatjara and English. In MAK Halliday (ed.) Cultural
Dynamics VI, 1-2 1993. Leiden, Netherlands: EJ Brill
Rose D., McInnes D., Korner H. 1992. Write it Right: Scientific Literacy in Industry. Sydney: Disadvantaged Schools
Program, NSW Education Department.
Trudinger, R. M. 1943. Grammar of the Pitjantjatjara Dialect, Central Australia. Oceania 13.3: 205-23.
Shore, Susanna. 1992. Aspects of a Systemic-Functional Grammar of Finnish. Unpublished doctoral dissertation.
School of English, Linguistics and Media. Sydney: Macquarie University.
Whorf, B.L. 1938. Some verbal categories of Hopi. Language 14: 275-286. (reprinted Language, Thought and Reality:
selected writing of Benjamin Lee Whorf, edited and introduced by J. B. Carroll. Cambridge Mass: MIT Press.)
PITJANTJATJARA PROCESSES
32
Whorf, B.L. 1950. An American Indian Model of the Universe. IJAL, 16: 67-72. (reprinted Language, Thought and
Reality: selected writing of Benjamin Lee Whorf, edited and introduced by J. B. Carroll. Cambridge Mass: MIT
Press.)
Wurm, SA 1971. Classification of Australian Languages, including Tasmanian. Current Trends in Linguistics, edited
by Thomas A Sebeok. Mouton: The Hague.