Missing image file: EWP2011-003.jpg

Morphosyntactic Correlates of Reference in Auye

Linguistic form and narrative role in a Papuan language

Michael Moxness

SIL International®

SIL Electronic Working Paper 2011-003, June 2011
Copyright © 2011 Michael Moxness and SIL International®
All rights reserved


Auye [1] employs a distinct reference strategy for narrative discourse. A “dominant” referent (or group) is consistently marked throughout the narrative with morphosyntactic forms that distinguish it from other referents. Languages around the world that have such a reference strategy usually mark the distinction on the referent np. Auye, instead, expresses the distinction by means that are both morphological (tense-aspect marking on the mainline verb) and syntactic (special temporal adverbial clause types). After a description of the reference strategy, a proposed Dominance Hierarchy will attempt to show the correlation between the relative degree of dominance assigned to the referents and their morphosyntactic expressions.

1 Introduction

This paper attempts to describe the rather intricate reference strategy employed in Auye narrative discourse. To understand the ways in which this system appears to be unique among the world’s languages, we will first take a brief look at what characterizes the different types of reference strategies.

Dooley and Levinsohn (2001:117ff.) discuss the two main types of reference strategies, the most common of which is the “sequential” (or “look back”) strategy. The well-known cross-linguistic study documented in Givón (1983) focused on this strategy and shed much light on the correlation between the activation status of referents and the linguistic devices that encode them, as expressed by the Scales in the Coding of Topic Accessibility and the underlying Iconicity Principle (1983:17-18). For discussion regarding the ways in which subsequent research has interacted with the results of this study, see Runge (2006:43ff.).

The other, much less common, strategy of reference has been termed “thematic” by Grimes (1978) and more recently a “VIP” (Very Important Participant) strategy by Dooley and Levinsohn (2001:119ff.). According to Grimes, the basic characteristic of this strategy is the managing of reference “in terms of a thematic policy, in which one referent is distinguished from the rest when introduced, and a special set of terms refer to it no matter how many other things have been mentioned more recently” (1978:viii).

The linguistic devices used by languages employing the thematic (VIP) strategy reveal some variation, but the apparent thread running through all of them is the distinct encoding of the np to indicate the relative salience of a referent. “The VIP marker can be a specific linguistic sign attached to the participant’s referring expression, or simply the use of underspecification” Runge (2006:43).

Before discussing the specific strategy employed in Auye, we will look at the thematic reference strategies of two Papuan languages.

2 Thematic reference strategies in Papuan languages

This section will present a quick overview of the findings of two published studies, each describing the thematic reference strategy of a Papuan language.

The strategy in Sikaritai (Lakes Plain, Tariku, East) relies heavily on two particles, and , which are “most directly connected with establishing and maintaining the relative dominance of characters and propositions throughout the discourse” Martin (1986:207). The particle occurs as a suffix on an np at or near the beginning of a narrative, which then “establishes the referent(s) as dominant participants, i.e., as initiators of the events that advance the story plot” (ibid., 208). Subsequent reference can be either np + or alone as a pronoun (used anaphorically). The particle is aligned with non-dominant referents, who are the participants and props that do not advance the plot, and who are dominated by the major participants. The particle can also function alone as a third person pronoun. A reversal in the dominance status of a referent is accomplished by the combination of the two particles as babí.

The data for Galela (West Papuan, North Halmahera) are presented in terms of “participant topicality” rather than the notion of dominance. The topical participant is always marked by a pronominal (“co-referential”) verb prefix, whereas non-topical participants are not. This prefixing system indicates the syntactic role (ten forms each for subject and object) of the referent in the clause Sheldon (1986:234-236). Topicality can be indicated at any given point in the narrative, and on occasion a minor participant can become topical through significant interaction with a major topical participant (ibid., 246).

These studies confirm some of the common means of referencing in those languages that employ thematic reference strategies. These include marking of the np, special particles and distinct pronominal forms (e.g., Galela’s co-referential verb prefixes).

3 Reference strategy in Auye

As stated above, languages employing the thematic (VIP) strategy normally rely on np marking or distinct pronominal forms to indicate the saliency of a referent. In contrast, Auye accomplishes this through discrete tense-aspect marking on the verb and the special use of temporal adverbial clause forms. What makes this particular strategy unique is that these linguistic devices operate independently of the particular encoding of the referent np. In other words, the distinctive morphosyntactic marking of this thematic strategy runs parallel to a fully operational, sequential strategy of reference.

Since these linguistic devices in Auye will be discussed in terms of a referent’s relative degree of assigned “dominance,” it is necessary to explain how this concept is to be understood for Auye narrative in this paper. A basic definition for the moment (it will be slightly revised later) is given in (1):

Referents marked as dominant are to be viewed by the addressee as the most salient in terms of the goals and outcomes of the narrative on a global (text) level.

The remaining referents are marked as non-dominant. These not only include minor characters and inanimate objects, but also antagonist, a main character who normally exhibits a high degree of interaction with other main characters. The significance of a main character being marked in a similar fashion to lesser characters will be discussed under 5.2.

The two sections below present examples of the dominant and non-dominant forms operative in the morphology of mainline verbs and in temporal adverbial clause types.

3.1 Verb morphology

The basic dominant mainline verb aspect suffixes are -taa imperfective, -yu perfective, and -Ø. [2] (Abbreviations are listed here at the end of the paper.)

The forms -taa imperfective and -yu perfective are complementary in that each can occur with other aspect markers and directionals, but not the same ones.[3] By far the most common occurrence of -taa is immediately following the verb without any intervening aspectual marking (2a); cf. also the text Unruly:11. In contrast, -yu (2b) can never occur in this position in mainline verbs; cf. also Unruly:3.

Okoo yuwa okoo me ee kaa taa pee-taa.
3pl det.pl 3pl poss house obl along go-ipfv
‘They set off for their houses.’
Kuwa bedo wiya ekea paaya pene-ata ayaa yiba daka-me-awee-yu.
dem.pl bird two 2d both grab-seq net.bag inside hit-if-insert-pfv
‘(He) grabbed both those birds, then killed (them) and put (them) in the net bag.’

A mainline form that is used mostly for dominant referents is a verb root followed by -Ø (3); cf. also Python:8. The class of punctual verbs that allows this is quite small but they occur in highly perfective contexts, such as points of high tension, especially discourse peak (3). Most verbs cannot occur in this form, but must take tense-aspect marking, such as for the verb pee ‘go’ in (2a) and (5a).

Kopaa mumo to kaa kaawaa wiya toko-Ø.
club head det.sg obl times two hit-Ø
‘(With a) club (we) hit on (the python’s) head twice.’

In (4) are examples of plural action[4] aspect (4a) Unruly:15 and directional (4b) suffixes followed by -Ø; cf. also Unruly:5. These contexts are usually much more imperfective.

...inii nakaayoo ee epoo bagee yuwa too ome-gee-Ø.
...1pl here house own people det.pl only sleep-rep-Ø
‘...only we home owners here stayed on.’
Uudee kaa abaga-yamoo-Ø.
river obl chase-downward-Ø
‘(He) kept chasing (them) down to the river.’

The non-dominant equivalent of -taa and -yu for mainline verbs is the tense-aspect form -g immediate (5a). The form used to indicate plural action is -k present plural action (5b).

Piya uguu bade-doo-(ta) pee-g-i.
wood sharp extract-dep-(seq) go-imm-3m
‘He extracted the sharp stick then ran.’
Menaamee gapine-taa-k-i.
younger.sibling writhe-dur-ppla-3m
‘The younger brother was writhing in pain.’
Toomoo daba pee-g-i.
sleepy freely go-imm-3m
‘The sleepiness went away.’

Verbs expressing human experience such as tiredness (5c), sickness, pain, dying, (12), as well as actions such as cutting self and stepping on a sharp stick, are marked non-dominant. Most of these expressions show subject agreement with the referent undergoing the experience.

An example of both dominant and non-dominant forms in the same sentence (as well as occurring with the same predicate) is shown in (6). A similar, non-narrative example is shown in (47).

Api yuwa kene muni-komee-g-ea,
woman det.pl door close-down-imm-3pl
amaa yiba kaama inii miya kene muni-komee-yu.
house.type inside from 1pl also door close-down-pfv
‘The women closed (their) door, (and) from inside the amaa house we also closed (our door).’

The occurrence of these forms in non-narrative genre and in speech will be briefly discussed under Appendix B. The discussion under Appendix C attempts to show the range of perfectivity inherent in the mainline tense-aspect forms for both dominant and non-dominant contexts.[5]

3.2 Temporal adverbial clauses

Six major types of temporal adverbial clauses will be discussed and exemplified here.[6] Three of the types are used exclusively for dominant referents when occurring as clause subject, and the other three are reserved primarily for use by non-dominant referents. This will be explained further, along with frequency distribution data, under 5.1.

The first type (7) is used exclusively for dominant referents as clause subject. The distinct features of this clause type are the morphemes too simultaneous and ko point of departure (similar to “topic” in other descriptions) (7a). These follow what appears to be a finite clause, but in actuality the clause has a neutralized form of subject person and number marking. Though either of the subject 3rd singular forms (-a 3f or -i 3m) can occur here, no person or number is referenced, as is shown by the plural subject in (7b).

Dee-yamaa-g-i too ko maga da kaa oma wiya to-p-a.
look-downward-imm-3n sim pod ground det.sg obl stone two exist-rpst-3f
‘When I looked down, there were two stones.’
Toomoo ome-taa-k-i too ko ee dekee-dakaa-k-a.
sleepy sleep-dur-ppla-3n sim pod house sound-red:sound-ppla-3f
‘While (they) were sleeping, (something) was scraping the house.’

The second type is also reserved exclusively for dominant referent use (8). The distinct features of the clause are the aspect marker -yaa habitual followed by the adverbial marker naaki (8a). The combination of the two elements is often used when the attempted action has a variable or unpredictable outcome, which is expressed in the main clause. The plural action form takes the auxiliary ta/se ‘do’ (8b) Unruly:7.

Yuumaane-maa-ta dee-yamaa-yaa naaki kuwa U. yuwa paka-taa-k-ea.
post.guard-comp-seq look-downward-hab adv dem.pldem.pl U. det.pl ascend-dur-ppla-3pl
‘When (they) posted guard (and) looked, those U. (warriors) were ascending.’
...koma kobe-taa si-yaa naaki yamo-ki-taa D. noonaa ma pakaa-maa-g-ea.
...canoe chop-dur aux.do-hab adv downward-ir-along D. some and ascend-comp-imm-3pl
‘...while (we) were chopping out a canoe, some other Dani people arrived.’

The third type, also for dominant referents, occurs with zero marking on the verb (9) Python:5. The number of punctual verbs in this class is quite small.

...amaa miyoo taa peto esee-waki-Ø ko manta kutuu da dimi nagi.
...house.type under along torch descend-shine-Ø pod python coil det.sg mind astonish
‘...when (I) shone the torch under the house, (there was) a huge, coiled python.’

The two forms in (10) are not used exclusively for dominant referents, though their distribution is still restricted, as will be shown in 5.1. The clauses only occur with the immediate tense-aspect marker -g, and the markers ko point of departure (10a) and naaki adverbial (10b) follow; cf. also Enemy:8, Unruly:17 and (29). The clauses are fully finite, unlike those in (7).

Ega-taa-g-i ko ee bedimi, bugi bedimi.
count-exh-imm-3m pod house six garden six
‘When he counted, (there were) six houses and six gardens.’
Okaa da doo-komee-g-i naaki magaa da kaa widigi taa.
bow det.sg carry-down-imm-3m adv ground det.sg obl touch not
‘When he reached down using the bow, (it) didn’t touch the ground.’

The next type is actually a relative clause followed by the generic noun for time, kaboo (11); cf. also Python:13 and Python:18. This is the form used for switching from a non-dominant referent in a previous clause to a subject of greater rank in terms of dominance in the following main clause, as will be discussed under 5.2 below.

...koogo toomuu-taa-k-a too ko bedo me-ani-g-i kaboo da wogee-yamoo...
quietly rest-dur-ppla-3n sim pod bird come-unc-imm-3m time det.sg shoot-downward
‘...while (we) waited quietly, when a bird came, (we) shot (it)...’

The last type (12) is used exclusively for non-dominant subject referents in both the adverbial and main clauses. It is very similar in form to the second clause type above (8), but here the only verb used is gaa ‘think, be’. Other examples of this type are shown in (13) and Unruly:20.

Daa-taga paa-taga eke-ase-g-ea gaa-yaa naaki mee otoogiyoo da boo-g-a.
cut-exh.pla red:cut-exh.pla leave-dec-imm-3pl be-hab adv dem.sg snake det.sg die-imm-3f
‘They cut [it] up and left [it], at which time that snake was dead.’

There were several interesting examples in the corpus showing the embedding of temporal adverbial clauses, such as (13); cf. also (11). Notice how the dominant clause type in brackets is embedded in the non-dominant clause.

Paa dekaa to naa segaa-(ta) pee-g-a gaa-yaa naaki,
spirit.being rain det.sg conjec sprinkle-(seq) go-imm-3f be-hab adv
[ega-ani-i-no taa dee-yamoo-Ø ko] yuma, U., U. ebo etaa.
[time-unc-irr-inf not see-downward-Ø pod] earlier U. U. big path
‘When it sprinkled (with the sun shining--a bad omen), [shortly thereafter when (they) looked,] (in accordance with the bad dream--another bad omen) earlier, (there were) U. (warriors), a huge group of U. (warriors were coming).’

The distinct distribution patterns of these six types of temporal adverbial clauses and their significance for Auye narrative discourse will be discussed under 5.

4 Modifications to the basic thematic strategy in Auye

Two situations have been found in the corpus that require a slight modification to our basic understanding of the thematic strategy used in Auye.

4.1 A referent can have both markings in the same text

A referent can have both dominant and non-dominant markings in the same text, depending on the referent’s role at any particular moment. When acting in a role considered either to be part of the dominant group (or in distinction with that of antagonist or others), the referent takes dominant marking (14a). When temporarily in a role that is distinct from that of the dominant one (i.e., both are doing separate actions), then the referent takes non-dominant marking (14b).

Anii ma, O. boo-ta ma, M. boo-ta ma, inii too obee-taa.
1sg and O. die-sta and M. die-sta and 1pl only follow-ipfv
‘I, the late O., and the late M., only we followed.’
Piya gee da yiba taa pe-sekee-(a)ni-g-i, kaboo da taa O. boo-ta goo-(a)ni-g-i.
tree root dem.sg inside along go-stay-unc-imm-3m time det.sg along O. die-sta creep-unc-imm-3m
‘(The pig) went and hid among the roots of a tree, at which time the late O. crept up on (it).’

Here is an example of both forms occurring in the same sentence (15):

S. pakaa-maa-g-i, epo ya-gape-komee-yu.
S. ascend-comp-imm-3m together dat:3.exc-down-pfv
S. arrived (and) together we baked (the food) for (other people).’

The two examples above reveal how a referent can receive dominant marking for actions done together with narrator in first person narrative. The example below is taken from a 3rd person narrative which regularly marks protagonist with dominant marking throughout the text (16a), similar to the marking of narrator in 1st person text. In this text, however, there is an instance of author intrusion, which forces a temporary, non-dominant marking of the author’s parents (16b).

Ekea yaai, amaa yiba pakaa-sekee-(a)n(i)-ta kene muni-ta-yu.
2d afraid house.type inside ascend-stay-unc-seq door close-exh-pfv
They were afraid (and) went into the house then closed the door.’
Anii ko peedi yoka ‬“Yegee ta-a-nak-i ye?” naadi
1sg pod small reason cry aux.do-irr-immin-3m ynq grn
koogo na-poda-maa-g-ea, ama nee-g-ea.
quiet obj:1sg-wrap-comp-imm-3pl breast dat.1sg:give-imm-3pl
‘Because I was small (they) wondered if I was going to cry, so they quietly held me (and) nursed me.’
4.2 The dominant referent is not identical with the “central” character

The data presented up to this point have been from texts in which the central character is marked as the dominant one. However, one “numskull” text that wasn’t included in the corpus for this study shows that this is not a fundamental association in Auye. Numskull is clearly the central character, yet takes non-dominant marking (17a). Instead, it is the unnamed, minor characters in (17b), who take dominant marking.

Manimaa dabaa wa too amo-ki-taa pe-seomi-g-i. Too mee-maa-g-i.
Manimaa scorned 3sg only over.there-ir-along go-depart-imm-3m long.time come-comp-imm-3m
Ipi daa-ge paa-ge, poda-awee-g-i.
taro.leaves cut-rep red:cut-rep wrap-insert-imm-3m
‘Dumb Manimaa, alone, goes out. After a while he returns, cuts up taro leaves, wraps (them and) puts (them) in (his net bag).’
Miimii wiya aa-iyage-ta-yu, ya-mone-ta-yu. Mapeda noonaa goo-manee-yu.
hunting.blind two rec-in.line-exh-pfv dat:3.exc-build-exc-pfv arrow some push-give.3dat-pfv
‘(They) made two bird hunting roosts in line with each other, one for (dumb Manimaa). (They) gave (him) some arrows.’

It appears, then, that dominance marking has a “suitability” feature associated with it. Those referents not deemed suitable for the role are demoted and assigned non-dominant marking.[7]

4.3 A modified understanding of the basic thematic strategy

We have now arrived at a fuller understanding of the basic thematic strategy of reference in Auye. A modified definition of dominance for Auye narrative as it relates to this strategy is as follows:

Referents marked as dominant are to be viewed by the addressee as the most salient and suitable in terms of the goals and outcomes of the narrative on both the local and global (text) level.

With this revised understanding of the constraints under which a narrative is crafted, we will now consider how this convergence of dominance and morphosyntax is worked out in the text corpus.

5 Text corpus analysis

The text corpus that served as the basis for analysis consisted of eighteen personal experience narratives authored by thirteen different individuals. The majority of the narratives can be classified as yaai-yaai menaa ‘scary story’ and were selected because they exhibited most of the features relevant to this study (e.g., multiple levels of dominance, identifiable antagonist). Travelogues were not included in the analysis due to the relative lack of these features . Since oral literature (folklore) presented its own set of problems (e.g., determining the perceived status of animals, or that of siblings who turn into spirit beings), they were also not included in this study.

Nine of the narratives are 1st person texts containing approximately 520 clauses and the remaining nine are 3rd person narratives containing approximately 470 clauses.

The frequency distributions displayed in the tables below have been confirmed by a more cursory analysis of similar texts from a larger corpus, but which were not included in this study.

5.1 Statistics according to morphosyntactic marking and role

The two tables in this section show the frequency distribution of subject verb marking in main clauses (19) and in temporal adverbial clauses (20) for 1st and 3rd person narrative.

In the tables, narrator encompasses references to self as subject in 1st person narrative and who receives the focus of attention on the global (text) level. Protagonist is the equivalent term for the individual (or group) occupying this role in 3rd person narrative. Associate signifies those who are normally introduced near the beginning of the narrative with the narrator or protagonist and share their activities and goals. Others are minor and often unnamed characters that have minimal interaction with the main characters, usually on the local level. Human experience (death, sensation), sub-human entities (animals, natural phenomena) and inanimate objects are also similarly marked and fall into this category. The term antagonist designates a main character that is characteristically at odds with the narrator or protagonist (e.g., enemy, spirit being, python).

Comments on the significance of the tabulations in (19) and (20) will be presented in 5.2. Special note is made of places where certain forms do not occur, not only in this corpus of selected texts, but in the larger corpus as well.

[A note regarding the verb types not included in this study: Main clauses having speech verbs (asi ‘say’, wega ‘speak’, tugi ‘invoke’) and the cognitive verb gaa ‘think’ were not included in the tabulations because, while dominant marking is maintained for narrator in 1st person narrative for these verbs, the distinction is neutralized for protagonist in 3rd person narrative (i.e., both dominant and non-dominant marking is allowed). No other verbs were found in the corpus that allowed this distinction between 1st and 3rd person narrative. Also, to avoid any skewing of results arising from the subject marking of reciprocals, the small number of verbs occurring with the reciprocal prefix aa- were not included.]

Table (19) shows the frequency distribution of the various types of subject marking on the verb in mainline clauses in relation to the role of the entity for both 1st and 3rd person narrative.

Frequency distribution of subject verb marking in main clauses for 1st and 3rd person narrative
1st person narrative texts3rd person narrative texts
V markingnarrator
(+ associate)
(+ associate)
-taa ipfv45 (6),(14a)Do not occur22 (2a),(2b)Do not occur
-yu pfv
V stem-Ø40 (3), (4a)02238 (4b), (32)014
-k ppla1*11 (14b)45 (6)25 (8a)4**6 (17a)18 (5b)57 (5a)
-g imm
* Two verbs, both expressing advancement in location, allow dominant referents to take finite verb marking. The text corpus for this study did not include travelogues, but one of these verbs, seome ‘proceed’, is readily found there.
** Two of the occurrences here are due to the special circumstance of author intrusion, as shown in (16b). The other two are verbs expressing advancement, as explained in the previous note (cf. Enemy:18).

In the table above, the tabulations for taa imperfective and yu perfective as well as -k ppla and -g imm were combined since their distinction was not relevant for this study. Discussion of the various morphological forms can be found under 3.1 and Appendix C. Links to examples of the forms in this paper and their distribution among the various roles as subject are also included in the table.

Table (20) shows the frequency distribution of the various types of temporal adverbial clauses in relation to the role of the entity as clause subject for both 1st and 3rd person narrative.

Frequency distribution of subject in temporal adverbial clause types for 1st and 3rd person narrative
1st person narrative texts3rd person narrative texts
Adv Cl typenarrator
(+ associate)
(+ associate)
FinCl + too ko13 (7a)Do not occur7 (7b)Do not occur
Vstem-yaa + naaki25 (8b)8 (8a)
V stem + ko3 (9)3 (13)
FinCl + koDo not occur24* (28)6 (10a)13* (29)
FinCl + naaki6 (10b)
FinCl + gaayaa naakiDoes not occur5* (12)Does not occur4* (13)
FinCl + kaboo0**24 (11)30024 (32)
* The subjects of both the adverbial and main clauses are members of these two groups (others and antagonist) only.
** There were two occurrences here of non-finite clauses (purpose and verbless/stative). Also, instances of “comma intonation” (14b) were not included in the tabulations since they are most likely a shortening of the fuller expression Kuwa kaboo... ‘At that time...’, which often starts a new sentence, and hence, expresses a lesser degree of integration of the head kaboo with the preceding clause. See Enemy:7.

The data from these two tables will reappear in more simplified tables under 5.2 as evidence for the possibility of a Dominance Hierarchy.

5.2 Is there a legitimate Dominance Hierarchy operative in Auye narrative discourse?

In light of the apparent correlation that exists between the relative dominance of a referent and its morphosyntactic expression, it is natural that we pursue the possibility of a Dominance Hierarchy for Auye narrative. The assumption that serves as a basis for all that follows is that narrator and protagonist receive the focus of attention on the global (text) level and therefore their occurrence in a narrative will correlate with the grammatical forms that express the highest degree of dominance.

The proposed Dominance Hierarchy is as follows:

   narrator  >  protagonist   >   associate  >  others  >  antagonist

The tables below attempt to provide evidence for the existence of the hierarchy. They are basically a reworking of the data presented in tables (19) and (20) above, except that the distinction between 1st and 3rd person narrative has been eliminated and the data combined. Only the data for narrator and protagonist have been kept separate in some tables.

The tables attempt to highlight the relative degree of dominance for adjacent pairs in the hierarchy. The tables also include a column on the far left that groups the markings (A, B, C, D) according to the apparent correlation between a referent’s relative dominance and its morphosyntactic expression. These groups are also hierarchically arranged:

   A  >  B  >  C  >  D

The first comparison to be made in terms of relative dominance is the larger one involving the entities on the far left side with those on the right (narrator, protagonist > associate alone, others, antagonist. The color highlighting in table (23) shows the dominance of narrator and protagonist in regard to the other entities.

Frequency distribution of subject verb marking in main clauses according to role and relative dominance grouping
Relative dominance
V markingnarrator / protagonist
(+ associate)
A-taa ipfv71Do not occur
-yu perf
V stem-Ø74536
B-k ppla5115676
-g imm

The table shows a clear preference on the part of narrator and protagonist for A group marking. This is strongly reinforced by the fact that taa imperfective and yu perfective do not even occur with the other entities.

The next comparison involves narrator > protagonist. Table (24) reveals the evidence for this ranking.

Frequency distribution of temporal adverbial clause types according to role and relative dominance grouping
Relative dominance
Adv Cl typenarrator
(+ associate)
(+ associate)
AReduced FinCl + too ko137Do not occur
V stem-yaa + naaki258
V stem + ko33
BFinCl + koDo not occur1257
FinCl + naaki
CFinCl + gaayaa naakiDoes not occur9
DFinCl + kaboo00213

Though the color highlighting for narrator (yellow) reveals the stronger preference for A group marking, the real evidence for the ranking narrator > protagonist lies in the fact that B group marking is allowed by protagonist (green) but not by narrator.

It should also be noted in this table that the relative dominance grouping A for temporal adverbial clauses supports the ranking (narrator, protagonist > associate alone, others, antagonist. As in table (23), group A marking does not occur with the other entities.

Tables (25) and (26) show the dominance of associate in relation to the entities to the right in the hierarchy (associate > others, antagonist). The figures reflect the total markings of associate as an entity, not according to role (i.e., including the instances of dominant marking when acting with the narrator and protagonist, not just the finite, non-dominant marking when acting alone.)

Table (25) gives evidence for the ranking (associate > others, antagonist in terms of subject verb marking.

Frequency distribution of subject verb marking according to role and relative dominance grouping
V markingnarratorprotagonistassociateothersantagonist
A-taa ipfv856028Do not occur
-yu pfv
B-k ppla1417149
-g imm

The figures in this table show the clear distinction in group A marking, again by the fact that taa imperfective and yu perfective do not occur with others and antagonist. Group B marking is also fairly distinct.

Table (26) gives evidence for the ranking associate > others, antagonist in terms of the subjects of temporal adverbial clause types.

Frequency distribution of subject in adverbial clauses according to relative dominance grouping and role
Adv Cl typenarratorprotagonistassociateothersantagonist
AReduced FinCl + too ko41188Do not occur
Vstem-yaa + naaki
V stem + ko
BFinite Cl + koDo not occur1237
Finite Cl + naaki
CFinite Cl + gaayaa naakiDoes not occur9
DFinCl + kaboo (RelCl)50913

As in the previous table, the percentage figures in (26) show the clear distinction in group A marking by the fact that the forms included there do not occur with others and antagonist. The lack of Group C marking for associate provides supporting evidence.

The last pairing in the hierarchy to examine is others > antagonist. Before presenting evidence for this ranking, we need to look at how dominance is expressed for Auye in general terms. The table in (26) above showed the correlation between A group marking and the more dominant roles of narrator, protagonist and associate. The table in (27) below shows that the subjects in these roles all express clear dominance vis-à-vis the subject of the main clause (indicated as > S in the table). There is never an instance where the subject of the adverbial clause is less in dominance than that of the main clause (though they may be the same subject or of equal status).

Frequency distribution in terms of the relative dominance of adverbial clause subject vis-à-vis main clause subject
Relative dominance
Adv Clause type> S of
main cl
Same S or =
to main cl
< S of
main cl
Aclause + too ko5350
V stem-yaa + naaki
Ø V marking + ko
BFinite Cl + ko194
Finite Cl + naaki
CFinite Cl + gaayaa naaki27
DFinCl + kaboo (Rel Cl)0411

What is noteworthy in the table is that the expression of relative dominance between the subject of the adverbial clause vis-à-vis the subject of the main clause extends to the lower groupings B and C as well. (Only the relative clause in D expresses adverbial clause dominance that is primarily less than or equal to that of the main clause, the significance of which will be discussed below.)

The following examples are limited to subjects in the roles of other and antagonist. The best example found in the corpus showing how relative dominance is expressed between these subjects actually also happens to come the closest to being a counter-example, specifically the first sentence in (28) Unruly:17.

Ese-ani-g-i ko T.J. ese-ani-g-i. Ese-ani-g-i ko mee mèe weda kaa...
descend-unc-imm-3m pod T.J. descend-unc-imm-3m descend-unc-imm-3m pod dem.sg man fear obl
‘When (the helicopter) landed, T.J. got out. When (he) got out, that man [the antagonist] in fear...’

Most likely, however, the helicopter here represents all the passengers in it, thereby exhibiting greater dominance than the one passenger who disembarks. The subject of the adverbial clause in the second sentence then expresses dominance in relation to the antagonist.

Similarly in (29), the spirit being that arrives on the scene and chases the other spirit being away is considered dominant by its occurrence in the adverbial clause. (Spirit beings normally occur in the role of antagonist, thus the one in the main clause is considered the greater antagonist.)

Mee paa da woo-ki-yoo esee-menaa-k-i ko paa taka ena yamoo.
dem.sg spirit.being det.sg upward-ir-place descend-vocalize-ppla-3m pod spirit.being moment one downward
‘When that spirit being came down (and) sounded, (the other) spirit being at once (fled) away.’

We are now in a position to examine the pairing others > antagonist. The table in (30) can be read as: “In regard to others and antagonist, the subject of the adverbial clause types listed on the left side do or do not express greater (>) or lesser (<) dominance vis-à-vis the subject of the main clause.”

Relative dominance of adverbial clause subjects vis-à-vis main clause subjects for others and antagonist *
Relative dominance
Adv Clause typeOthers S >
main cl S
Others S <
main cl S
antagonist S >
main cl S
antagonist S <
main cl S
Aclause + too koDo not occur
V stem-yaa + naaki
Ø V marking + ko
BFinite Cl + koyesnonono
Finite Cl + naaki
CFinite Cl + gaayaa naaki
DFinite Cl + kaboo (RelCl)yesyes
* Tabulations for some categories in the table were small, so texts outside the corpus were examined.

The table shows that only others, as the subject of a temporal adverbial clause, can exhibit dominance vis-à-vis the subject of the main clause. Antagonist does not allow this, which also provides evidence for its inclusion as a separate entity in the hierarchy.

The data above might lead to the conclusion that storytellers avoid expressing the dominance of antagonist over the entities higher in the hierarchy. However, much of the pre-resolution material of a text can portray antagonist in a very dominant position, and some texts actually have antagonist triumphant in the end. The author accomplishes this by the prolific use of the simple clause, e.g., ‘(The enemy) shot so-and-so’; cf. also Enemy:17. Another method involves demotion by the use of a relative clause, which expresses new information (31) Python:22-23.

Okooto pe-seome-'g-a yoka ega-a-no taa kaboo ko
mother go-proceed-rpst2-3f reason count-irr-inf not time pod
amoo ko kameno keo, [okooto patame-g-a] yoo.
over.there pod dog yelp [mother kill-imm-3f] place
‘The mother (dog) had left, so in a matter of moments, over yonder, a dog yelped, (from) the place where (the python) killed it.’

If antagonist does happen to appear as subject of an adverbial clause (and only in group C and D forms), it expresses either equal or lesser dominance than the subject of the main clause, never greater. The only adverbial clause type that allows this lesser dominance in the adverbial clause is D, the relative clause form, as shown in (27) above and exemplified in (32) and (11).

Mee-maa-g-ea kaboo da taa ko waa mapeda kaa taki.
come-comp-imm-3pl time det.sg along pod 3sg arrow obl cover
‘When (the enemy) arrived, he pounced on (his bow and) arrows.’
5.3 Summary of evidence

The table in (33) provides evidence for the ranking of entities within the proposed Dominance Hierarchy.

Summary of evidence for the ranking of entities within the Dominance Hierarchy
Relative rank in terms of dominanceMain evidence for the ranking
narrator > protagonistProtagonist can occur as subject in lower ranked adverbial clauses (B group in (26)) and can take non-dominant marking for verbs of speech and cognition. Narrator does not allow either.
narrator, protagonist > associateNarrator and protagonist take group A marking for temporal adverbial clauses, but associate does not when acting alone, as shown in table (24).
associate > othersAssociate is closely identified with narrator or protagonist and is almost always introduced near the beginning of the narrative. It can and often does share similar A group marking (26). Others are introduced later in the narrative and do not take A marking.
others > antagonistTable (30) shows that others, as the subject of an adverbial clause, can exhibit dominance vis-à-vis the subject of the main clause. This is not allowed for antagonist, which also provides evidence for its inclusion as a separate entity in the hierarchy.
6 Conclusion

Discourse studies are commonly perceived as addressing nebulous topics that can only be described in terms of “principles” or “tendencies”. Though this paper does not put forward fixed rules resembling those operating on the morphosyntactic levels, the study shows that the Auye storyteller is very aware of the grammatical forms that constrain the crafting of the narrative on the text level. The careful manipulation of the rich and varied forms that express the relative salience and dominance of the characters in a story makes storytelling in Auye an intricate work of art. [8]

Appendix A: Overview of Auye verb tense and aspect

The purpose of this appendix is to give an overview of the verb tense and aspect forms that are used in the paper. The listing is not exhaustive of the forms that occur in the language, but should give a good idea of how tense and aspect are used in narrative text. The examples below are shown in their everyday speech (i.e., non-dominant) forms.

The tense forms in (34) are listed for the verb me ‘come’ for 1st person subject -a. The irrealis marker -e is an allomorph suffixed to the verb stem in non-past contexts (34h-j).

a.me-'ta‘(I) came (long ago)’ far past
b.me-sek-a‘I came (long ago)’ personal past
c.me-'m-a‘I came (yesterday or before)’ past
d.me-'p-a‘I came (recently today)’ recent past
e.mee-maa-'g-a‘I arrived (recently today)’ recent Past 2 (following -maa completive)
f.me-g-a‘I am coming (now)’ immediate
g.me-taa-k-a‘I am in the process of coming’ present plural action (following -taa durative)
h.me-e-nak-a‘I am about to/intend to come’ immininent
i.me-e-p-a‘I will come (later today)’ near future
j.me-e-t-a‘I will come (tomorrow or beyond)’ future

Some of the forms listed above function as a combination of tense and aspect. For instance, the two tenses that reflect present time are distinguished according to whether the action is viewed as a whole (35a) or as a process, as when preceded by a plural action suffix such as -taa durative (35b).

a.aba-g-a‘I am searching’ immediate
b.aba-taa-k-a‘I am in the process of searching’ present plural action

Besides the fact that -g immediate functions as the (non-dominant) mainline tense-aspect, another indication that it is not merely an absolute tense is shown by its additional function of expressing new information. The utterance in (36) is expressed in terms of immediate relevance, though the event may have occurred several months previous. (The same form could function more as a tense, however, by referring to an event that just happened.)

Ni me yoga to boo-g-i.
1s poss child det.sg die-imm-3m
‘My child died.’

Another form that is a combination of tense and aspect is the suffix -'ta (34a) (uninflected for subject person-number marking) can function as far past tense (37a) or stative aspect (37b); cf. also (14a):

a.boo-'ta‘(He/she/etc) died (long ago)’ far past
b.boo-'ta‘(He/she/etc) is dead’ stative

The table in (38) shows tense as encoded in verbs of various degrees of perfectivity and syllable structure. The tense -sek personal past is rare so has not been included. Notice that the verbs of lowest inherent perfectivity (i.e., expressing plural action), such as yegee ‘cry’ in the far right column, take the auxiliary se/ta ‘do’ in non-present contexts.

Far Pastpee-'tano-'tadee-'taome-'tawetoge-'taseo-me-'tapene-ase-tayegee se-ta
Pastpee-m-ano-'m-adee-m-aome-m-awetoge-m-aseo-me-m-apene-ase-m-ayegee se-m-a
Recent Pastpee-'p-ano-'p-adee-'p-aome-'p-awetoge-'p-a-pene-ase-g-ayegee se-p-a
Recent Past 2-----seo-me-'g-a-
Present Plural Action-------yegee-k-a
Imminentpo-o-nak-ana-a-nak-adi-i-nak-aumi-i-nak-awetoge-a-nak-aseo-ma-a-nak-apene-asi-i-nak-ayegee ta-a-nak-a
Near Futurepo-o-p-ana-a-p-adi-i-p-aumi-i-p-awetoge-a-p-aseo-ma-a-p-apene-asi-i-p-ayegee ta-a-p-a
Futurepo-o-t-ana-a-t-adi-i-t-aumi-i-t-awetoge-a-t-aseo-ma-a-t-apene-asi-i-t-ayegee ta-a-t-a

The sets of verbal aspect morphemes that follow can be divided into three groups on a structural basis. The first group of aspectual forms (Aspect 1) in (39) can take tense-aspect (here -g immediate) and subject person-number marking. Several directionals also follow this pattern. In this context they express a relatively high degree of perfectivity. The forms here are in their irrealis expressions.

a.noo-maa-g-a‘I finished eating’ completive
b.noo-taa-g-a‘I ate (it) all up’ exhaustive
c.noo-gee-g-a‘I went, ate and returned’ go-do-return (directional)
d.noo-komee-g-a‘I swallowed (it)’ downward (directional)
e.doo-yamaa-g-a‘I carried (it) down’ downward (directional)
f.tamee-pe-ase-g-a‘I threw (it) away’ terminate
g.gade-ase-g-a‘I tied (it)’ decisive
h.gadi-(a)ni-g-i‘(The pig) got snared (by the rope)’ uncontrolled

The second group of aspectual forms (Aspect 2) in (40) also take tense-aspect (here -k present plural action) and subject person-number marking. The plural action that they encode expresses a low degree of perfectivity. See Appendix C for fuller discussion regarding perfectivity of these forms.

a.woge-taa-k-a‘I am in the process of shooting’ durative
b.woge-gee-k-a‘I am shooting repetitively (over a period of days/weeks)’ repetitive
c.woge-pegee-k-a‘I am shooting repetitively (as I go)’ repetitive 2
d.woge-ii-k-a‘I am shooting iteratively (one-after-another)’ iterative
e.woge-a woge-a-k-a‘I am shooting periodically (at short intervals)’ periodic
f.woge-naa-k-a‘I am going around shooting’ distributive
g.pee-daa-k-ea‘They are dispersing’ dispersive
h.ase-dee-k-a‘I am trying to say/ask’ attempt
i.woge-yamoo-k-a‘I am shooting downward (in some plural manner)’ downward (directional)
j.woge-tagaa-k-a‘I am shooting and hitting the mark (in some plural manner)’ exhaustive (plural action)

The third group in (41) does not take additional inflection.

a.pee-ta‘after going’ sequential
b.pee-yaa‘goes, would go’ habitual
c.po-o-kee‘customarily/rightfully go’ Unruly:21 customary

Other (non-aspectual) markers are listed in (42).

a.po-o-no‘to go’ infinitive
b.pee-yaa-g-a‘I caused to go’ causative
c.pee-s-i‘in order that (it) might go’ Python:10 purpose

Appendix B: Dominance in non-narrative discourse

This section will present a brief overview of the ways in which dominant forms are used in non-narrative discourse, specifically speech, procedural and descriptive discourse contexts.

The marking of dominant and non-dominant referents is neutralized for speech, including non-narrated, everyday conversation. The form for all referents is non-dominant (43):

“Anii weda-g-a yoka po-o-i!”
1sg afraid-imm-1sg reason go-irr-imp
“I’m afraid (of the attackers), so go!”

In procedural texts the forms -taa imperfective and -yu perfective are used in the very same way as they are in narrative.[9] The following excerpt from a text explaining the process of planting taro (44) shows how these forms function in mainline clauses:

Petamanii senaa keta dee-taa, dee-maa-yu naadi
first brush before clear-ipfv clear-comp-pfv then
piya kobe-taa, piya kobe-maa-yu, piya kobe-maa-ta
tree chop-ipfv tree chop-comp-pfv tree chop-comp-seq
gapuu daa-taa, gapuu daa-maa-yu naadi
undergrowth cut-ipfv weeds cut-comp-pfv then
iyoo we-taa, we-maa-yu, we-maa-ta...
stalk plant-ipfv plant-comp-pfv plant-comp-seq...
‘First the area is cleared. The area is finished being cleared, then the trees are cut down. The trees are finished being cut down, then after finishing cutting down the trees the undergrowth is cut. The undergrowth is finished being cut, then the stalks are planted, finished planting, then after finishing planting...’

Compare the above marking with the similar marking in the narrative text here (45):

Ne-ataagi ma, anii ma pee-doo-ta nepoubo muni-i-no see naadi
1sg.poss-father and 1sg and go-dep-seq tree.sp tend-irr-inf such grn
kabo daa-gee-ta keage piya kapa-maa-yu naadi
base cut-gdr-seq floor wood split-comp-pfv then
wai taga-taa, taa-gee-ta kaama ootoo da taa-gee-ta piya agiyoo sikii naadi
beam pick-ipfv pick-gdr-seq from ridge.pole det.sg pick-gdr-seq wood thing enough then
gekaa taga-taa, gekaa sikii naadi daa-gee-yu, uduma sikii naadi...
fern pick-ipfv fern enough then cut-gdr-pfv all enough then..
‘My father and I went to build a bird hunting roost, and after going to cut the base, then splitting the floor wood, we picked the beam (wood). After picking the wood, then after picking the ridge pole, the wood materials were enough so we picked the ferns. The ferns were enough so we finished cutting. Everything was done so...’

A descriptive text explaining the different forms of hunting in (46) shows the similarity in function of -taa and other imperfective aspect markers such as -yaa habitual and -Vkee customary.

Kameno maa-yaa mee miya eta gakataa se wugi-yaa,
dog put-hab dem.sg also path each neg kill-hab
kaagaa ena waa naa naagoo too wugi-yaa.
time one approx indef day only kill-hab
Mapeda me mee mèe idima se woge-taa kiya
arrow ag dem.sg person all neg kill-ipfv but
mèe ena ena too mapeda me yaa woge-ekee.
person one one only arrow ag wild.pig kill-cust
‘It’s the same for hunting with dogs. (The dog) doesn’t get game each time out; only every once in a while. As for using arrows, not everybody gets game, but usually only one (or so) person does using arrows.’

As was exemplified for narrative in (6), a non-narrative instance of both dominant and non-dominant forms in the same sentence (as well as occurring with similar predicates) is shown in (47).[10]

Kameno mee ko peedi yaato a me muni-taa si-yaa,
dog dem.sg pod small side 2sg ag care.for-ipfv aux.do-hab
ebo yaato wa me ka-muni-g-i see.
big side 3sg ag obj.2sg-care.for-imm-3m such
‘As for a dog, when [he] is young, you care for [him]; when he is big, he cares for (feeds) you.’

The above data lead to a discussion regarding the perfectivity of these aspectual forms, which is presented in Appendix C.

Appendix C: Perfectivity in Auye verbs

The material in this appendix will consider the degree of perfectivity inherent in Auye verbs used in narrative (and most non-narrative) discourse.

Table (48) shows the correlation between the inherent perfectivity of the verb form and its tense-aspect marking. The verb classes are arranged from most perfective (1) to most imperfective (3). Glosses for the particular aspect and directional suffixes needed to interpret the data are included in the two notes at the bottom of the table. (See Appendix A for comparisons of these tense and aspect forms according to their respective paradigms and distribution patterns.)

Degrees of perfectivity in mainline tense-aspect verb forms in Auye narrative
Finite formsNon-finite forms
+ -g imm
+ -k ppla
+ -yu perf
+ Ø
Degree of
Verb class+ Verb aspect*,
directional** suffix
1Punctual  verbs, e.g.
eke ‘leave’
pene ‘hold’
-asi punc (only)eke-asi-g-i
Do not occurDoes not occureke-Ø
2Most transitive and intransitive verbs, e.g.
to ‘stay’
pee ‘go’
no ‘eat’
daa ‘cut’
aba ‘search’
doo ‘carry’
-asp 1
-dir 1
-dir 2
Do not occur
Do not occurpee-taa-Ø
-dir 1, 2Do not occurdaa-komee-k-i
-asp 2pee-taa-k-i
-dir 1, 2 +/- se aux.dodaa-komee taa-k-i
doo-yamoo taa-k-i
-asp 2 +/- se aux.dodoo-naa se-gee-k-i
noo-gee taa-k-i
doo-naa se-gee-Ø
3Imperfective verbs, e.g.
yegee ‘cry’
toomuu ‘rest’
-asp 1
-dir 1
Do not occur
-asp 2Do not occuryegee-taa-k-i
Do not occuryegee-Ø-Ø
seaux.do-dir 2yegee se-amoo-k-iyegee se-amoo-Ø
* Aspect 1 suffixes occur in perfective contexts and include -maa comp and -taa exh. For finite forms they are followed by -g immediate and subject person-number marking. aspect 2 is a set of six plural action aspect suffixes, such as -gee repetitive, -naa distributive, as well as -taa imperfective/durative. For aspect 2 suffixes the auxiliary ta/se ‘do’ is obligatory preceding any subsequent inflections in non-present contexts. Reduplicated forms also follow this pattern. In this table the finite forms are presented using -k present plural action and subject person-number marking.
** Directional 1 suffixes occur in perfective contexts and include -komee downward. The directional 2 suffixes occur in both perfective and imperfective contexts and include - amaa/-amoo sidewards, -yamaa/-yamoo downward and -waa/-woo upward.

The top of the table shows the forms having the highest degree of perfectivity, specifically those verbs inflected with -asi punctual (meaning the action occurred only once), -maa completive and other aspect 1 and directional suffixes. These are then inflected with the tense-aspect form -g immediate. The remaining finite forms in the middle and lower part of the table are less finite, occurring with aspect 2 (plural action) markers and inflected with the corresponding -k present plural action tense-aspect suffix.

The data in the table, however, seem to show evidence for three levels of perfectivity, as highlighted in red in the table for the three forms of daa-komee ‘cutting or biting action’, for which only two non-finite (dominant) forms are marked, and for the three forms of daa-yamoo ‘carry downward’, for which only one non-finite form is represented. Just as interesting is the fact that pee-Ø-g-i ‘he went’ (unmarked for aspect; highlighted in blue) happens to fall between the two forms of daakomee that are marked with the highest degrees of perfectivity. The crux of the matter lies in the fact that the corresponding dominant form is pee-taa-Ø. If this is the case, then we have the plural action aspect -taa occurring in a context that would be considered higher in terms of perfectivity than merely durative aspect.

The following examples will attempt to show that -taa encodes a range of perfectivity. First, support will be given for grouping -taa ipfv with the other aspect 2 morphemes. The example in (49) is taken from a text describing what is done to fruit bats after the tree in which they are roosting is felled:

Noonaa daka-me-taa, noonaa mapeda me woge-eii.
some hit-if-ipfv some arrow ag shoot-iter
‘Some are beaten, some are shot with arrows (one after another).’

Next, the following group of examples show how the aspect 2 (plural action) suffixes interact with the auxiliary ta/se ‘do’. The two examples in (50) show that the aspects can occur serialized, each suffixed to the auxiliary, with the final form carrying the finite marking.

...ikii me pekoo-pekoo, pakaa-pakaa se-naa se-gee-k-ea mee.
...2pl ag lazy-lazy red.lazy-red.lazy aux.do-dist aux.do-rep-ppla-3pl dem.sg
‘...your continual going around being lazy...’
…se ka-doo-ii se-gee ta-a-p-a.
neg dat:2sg.exc-carry-iter aux.do-rep aux.do-irr-nfut-1sg
‘...I won’t keep bringing [food] for you.’

The three examples in (51) show -taa following the same pattern as the other aspect 2 suffixes in (50), but occurring without the auxiliary ta/se ‘do’. The most likely reason for this is fact that -taa is the most generic of the plural action suffixes due to its expression of the very common durative aspect. It seems reasonable, therefore, that the auxiliary ‘do’ drops out in this context. (It is also noteworthy that -taa is the only one of the plural action suffixes that does not take accent, just as auxiliary ta/se ‘do’ does not.)

...bugii da yaa me noo-gee Ø-taa-k-ea...
..garden det.sg wild.pig ag eat-rep Ø-dur-ppla-3pl...
‘...wild pigs are eating the garden (over a time period of several days)...’
...yaa ena ese-taa paka-taa Ø-taa-k-i.
...wild.pig one descend-ipfv ascend-ipfv Ø-dur-ppla-3m
‘...a wild pig was continually going up and down.’
Yoo da kaa kiyoo agiyoo gape-taa no-taa Ø-taa see yoo.
outside det.sg obl there thing bake-ipfv eat-ipfv Ø-ipfv such place
‘Outside there is the regular place for baking and eating.’

Another factor to consider is that the form does not appear at the points of highest tension (e.g. discourse peak), where forms expressing higher perfectivity normally occur.

However, evidence that the form -taa also represents actions that are more perfective exists in the marking of speech verbs, which provides a controlled context where a legitimate comparison of dominant and non-dominant usage can be made. In 1st person narrative, speech verbs always have dominant marking for narrator as subject (see Python:4 and Python:24). The non-dominant marking, however, is -g imm Unruly:21.

Evidence for -g immediate encoding a range of perfectivity in non-dominant contexts also exists. Consider the following examples, the first of which would be the appropriate question posed to someone in the process of searching (52a). The statement in the second example (52b) also expresses an action in process.

Aa maagiyoo aba-g-e?
2sg what search-imm-2sg
‘What are you looking for?’
Api wiya me-g-ea.
woman two come-imm-3pl
‘Two women are coming (in sight, up the airstrip).’

In light of the data presented above, along with the data from non-narrative discourse in Appendix B (the imperfective usage of the forms as compared in (47) is especially revealing in terms of the discussion here), the conclusion to be drawn appears to be that both -g and -taa encode a range of perfectivity in their respective contexts. The following table (53) attempts to convey the comparative ranges of dominant and non-dominant marking of perfectivity in Auye verbs. (Not included for consideration is -Ø, which may occur in both perfective and imperfective contexts, as seen in (48) above.)

Degree of perfectivitynon-dominantdominant
high-g immediate-yu perfective
--taa imperfective
low-k present plural action
Appendix D: Texts
Python Attack
M. ma, E. ma inii wedo ome-taa-k-i too,
M. and E. and 1pl three sleep-ipfv-ppla-3n sim
M., E. (and I), we three, while sleeping,
deka mee ko maamaa we-taa-k-a-o see naagoo,
rain dem.sg pod much rain-ipfv-ppla-3f-emph such day
on a day the rain, it was really falling,
daatooyiba yupigapi kameno keo.
night middle dog yelp
in the middle of the night (we heard) a dog yelp.
“Mee ko otoogiyoo” ase-taa.
dem.sg pod snake say-ipfv
“That’s a snake,” I said.
Peto wake-ata kaama amaa miyoo taa peto esee-waki ko
torch shine-seq then house.type under along torch descend-shine pod
When after (I) shined the torch and then shined under the house
manta kutuu da dimi nagi.
python coil det.sg mind astonish
(there was) a huge python coiled.
Poga yaato tapa kaa amo-kiyoo to-p-a.
tail part open.area obl sideward-place stay-rpst-3f
The tail end was over in an open place.
Kuwa poga yaato kuwa yaka me egaa wedo kobe.
dem.pl tail part dem.pl axe ag count three chop
That tail end (I) chopped three times with an axe.
Boodi kaama idima yigi-me-gee-ta ee yiba me-see-me-g-a.
pain from all collect-if-gdr-seq house inside come-enter-if-imm-3f
From pain it curled up and then came inside (under) the house.
Pee-s-i naadi keage see-taa-ta kaama motaga gakee yamo-kiyoo mumo to kaa.
go-purp-3n grn floor separate-exh-seq then tree.sp plank downward-place head det.sg obl
To make it go away (I) separated the floor (slats) and then (hit it) on the head with an ironwood plank.
Mumo to kaa asi-yoo naku naku see egaa wui to waa naa,
head det.sg obl upward-place strike strike such count four det.sg approx conjec
(I) struck it on top of the head about four times,
boodi kaama eke, amoo kene-awee pee-g-a.
pain then leave sideward door-through go-imm-3f
(It was) in pain then (I) stopped (and) it went through the door.
Kene-awee pee-g-a kaboo
door-through go-imm-3f time
When it went through the door,
mumo to kaa kopaa daka-ma-a-no see naadi
head det.sg obl club hit-if-irr-inf such grn
in order to club (it) on the head,
“Peto maamaa na-wake-asi-i” asi-yaa naaki
torch much dat:1sg.exc-shine-dec-irr say-hab adv
when (I) said, “Shine lots of light for me,”
peto katu kee-goo-'g-a, peto ago po-o-no naka.
torch short aux.be-tran-rpst2-3f torch light go-irr-inf unable
the torch had been (burned) short, it wouldn’t light.
Otoogiyoo da taka ena amoo, kandang amoo pee-g-a.
snake det.sg moment one sideward pen sideward go-imm-3f
In a moment the snake was over yonder, beyond the pen.
Peto peenaa ago ago se-(e)pe-ase-g-a kaboo kopaa egaa wiya.
torch little light light aux.do-term-dec-imm-3f time club count two
When a little light was going, (I) clubbed (it) twice.
Amu-yaato boo yupi ki-see ki yiibaga pee-g-a.
sideward-side grass middle ir-manner immediately quickly go-imm-3f
It immediately went quickly over through the grass.
Kameno okooto pe-seo-me-'g-a yoka
dog mother go-depart-if-rpst2-3f reason
The mother dog had left, so
yoga-mee ee yiba boo-p-a, digiyoo kameno, pooduu to boo-p-a.
child-kin.sg house inside die-rpst-3f black dog young det.sg die-rpst-3f
the puppy inside the house was dead, a black dog, the offspring, was dead.
Okooto pe-seo-me-'g-a yoka ega-a-no taa kaboo ko
mother go-depart-if-rpst2-3f reason count-irr-inf not time pod
The mother dog had left, so not long thereafter,
amoo ko kameno keo, okooto pata-me-g-a yoo.
sideward pod dog yelp mother kill-if-imm-3f place
over yonder a dog yelped, (from) the place where (the python) killed the mother.
Ae, kaasee agiyoo? weda kaa “Mèe taba me-e-i!” ase-taa.
Oh why thing fear obl people group come-irr-imp say-ipfv
Oh my, what’s happening? (and) in fear I said, “Everybody come!”
Api, paayogaa idima mee-maa-g-ea.
woman man all come-comp-imm-3pl
Women, men, everyone arrived.
Peto wake-ata okooto keo keo se-p-a yoo pee-maa-g-i too ko
torch shine-seq mother yelp yelp aux.do-rpst-3f place go-comp-imm-3n sim pod
After shining a torch (and) upon arriving at the place where the dog had been yelping,
otoogiyoo da pe-seo-me-'g-a, kameno to boo-p-a.
snake det.sg go-depart-if-rpst2-3f dog det.sg die-rpst-3f
the snake had gone, the dog was dead.
Kagoo ebo, mee otoogiyoo yamo-kitaa pee-ta, ki-see to-p-a.
anger big det.sg snake downward-along go-fpst ir-manner stay-rpst-3f
(I) was very angry, the snake took off down (somewhere), and that’s how it ended.

Enemy Attack
Ne-ataagi waa peedi kaa U. me woge-tagaa se-ta menaa.
1sg.poss-father 3sg small obl U. ag shoot-exh:pla aux.do-fpst speech
(This is) the story of an enemy warrior attack in my father’s youth.
Wa-awaagi nokaa yuwa ma okoo mèe eebaa pee-doo-ta bugi kaa.
3sg.poss-older.brother group det.pl and 3pl people many go-dep-seq garden obl
He and his older brothers, they were many in all, went to the garden.
Pee-maa-ta yaga-anii yuwa kabata goko-me-taa-k-ea.
go-comp-seq older.brother-kin.pl det.pl treeless.area clear.brush-if-dur-ppla-3pl
After arriving, the older brothers were clearing brush in a treeless area.
Ne-ataagi waa motoo naa kobe-gee-ta damee tagaa to kaa pe-wataga-an(i)-ta,
1sg.poss-father 3sg tree.sp conjec chop-gdr-seq tree.sp branch.base det.sg obl go-lean-unc-fpst
My father, a motoo tree (had earlier been) cut down and was leaning on a damee tree,
mee damee tagaa to kaa pe-wataga-an(i)-ta yoka motoo koto naadi kaama
dem.sg tree.sp branch.base det.sg obl go-lean-unc-fpst reason tree.sp bridge grn then
and since it was leaning on the damee tree, it bridged the two, so
damee tagaa to kaa pe-seke, neataagi waa, wa me okaa mapeda yuwa ma.
tree.sp branch.base det.sg obl go-reach 1sg.poss-father 3sg poss bow arrow det.pl and
he climbed up into the damee tree, my father, he and his bow and arrows.
Kuwa kaboo da taa ko yaga-anii yuwa me menaa,
dem.pl time det.sg along pod older.brother-kin.pl det.pl poss sound
At that time, there was the sound of the older brothers working,
namo iye wugu dapuu naadi de-amaa-g-ea ko, yoga-anii yuwa,
taro leaf rustle red:rustle grn look-sideward-imm-3pl pod older.brother-kin.pl det.pl
there was rustling of the taro leaves so when they looked, the older brothers,
U. okaa goe-woo-k-ea, okaa kimuu.
U. bow pull-upward-ppla-3pl bow bent.like
U. warriors were pulling back their bows, the bows were bent.
Okaa goe-woo-k-ea yoka au-yii, pee-daa.
bow pull-upward-ppla-3pl reason yikes-exclamation go-dist
They were pulling back their bows, so (they) dispersed in terror.
Yaga-anii ena ki-taa, ena ki-taa.
older.brother-kin.pl one ir-along one ir-along
The older brothers went one here, one there.
Ne-ataagi waa au naadi, motoo koto naadi
1sg.poss-father 3sg yikes grn tree.sp bridge grn
My father was in terror so, using the felled motoo tree,
yuubaga magaa da kaa kagaa-ta ese-an(i)-taa.
quickly ground det.sg obl jump-seq descend-unc-ipfv
he quickly jumped to the ground.
Kagaa-ta kaama namo yupii yuubaga pee-taa-k-i kaa
jump-seq then taro middle quickly go-dur-ppla-3m obl
After jumping, in his run through the taro,
egamaa otoo to kaa ki-yoo wego otoo to kaa tani me-muguu-pee-g-a.
buttocks stump det.sg obl ir-place hip stump det.sg obl pain come-advance-go-imm-3f
in the area around his buttocks, near the hip, was a shot of pain.
Emaa woge-taa-g-i too ko komaakede.
eye shoot-exh-imm-3n sim pod arrow.type
When he looked, (there was a) komaakede arrow.
Ne-ataagi wa kaa woge-taa-g-ea.
1sg.poss-father 3sg obl shoot-exh-imm-3pl
My father was hit.
Bade-gee-ta kaama ka-pee-taa-k-i yuubaga, amo-ki-yoo naa kaa de-amaa-g-i ko
extract-gdr-seq then inc-go-dur-ppla-3sg quickly sideward-ir-place conjec obl look-sideward-imm-3sg pod
After extracting (it), he quickly advanced (and) when he looked over yonder,
ena, ne-ataagi waa amo-ke-taa, mapeda.
one 1sg.poss-father 3sg sideward-ir-along arrow
(there was) one. My father went over (and) shot (him).
Mee U. paa to woge-ta-yu ...
dem.sg U. fellow det.sg shoot-exh-pfv
(He) shot that U. fellow ...
Unruly Visitor
... Ega ena naagoo “Inii kaasee ukuwa-a-p-ea, maasee ukuwa-a-p-ea?”
... count one day 1pl how do-irr-nfut-3pl red:how do-irr-nfut-3pl
... Another day (we wondered), “What are we going to do?”
‘Inii too E. ma inii kuwa idima paka-doo-ta A. dee kaa’
1pl only E. and 1pl dem.pl all ascend-dep-seq A. river obl
Just us, E. and all of us went up the A. river and
keno piyaa ena kobe-maa-yu.
tree.sp tree one chop-comp-pfv
cut down a kenoo tree [a canoe wood].
Mee kenoo toga-agoo da, mee eke.
dem.sg tree.sp hollow-have det.sg det.sg leave
That kenoo was hollow (so) left it.
Ki-yoo kaama ese-doo-ta D. piya ena yapeo kobe-amoo.
ir-place then descend-dep-seq D. tree one tree.sp chop-sideward
From there (we) went down to D. and chopped down a yapeo tree.
Anii pee-doo-ta A. dee kaa,
1sg go-dep-seq A. river obl
I went to the A. river,
A. dee kaama uduma koma kobe-taa si-yaa naaki
A. river from all canoe chop-ipfv aux.do-hab adv
and from A. river, as we all were making a canoe,
yamo-ki-taa D. noo-naa ma pakaa-maa-g-ea.
downward-ir-along D. some-conjec and ascend-comp-imm-3pl
from down below some D. people and others arrived.
“Koma se kubi-i-i” see wee-g-ea, see ukuwa-ta mee ko
canoe neg chop-irr-imp such speak-imm-3pl such do-sta dem.sg pod
“Don’t make a canoe,” they said, regarding which,
“TJ me koma me-yaa-p-i” see wega-g-ea.
TJ ag plane come-caus-nfut-3sg such speak-imm-3pl
“TJ is going to send a helicopter,” they said.
Mee kaama inii ede ke-taa.
dem.sg from 1pl happy aux.be-imm
We were happy at that (news).
Mee kaama inii me wedaa-p-e dimii siki.
dem.sg from 1pl poss afraid-rpst-1pl.exc mind stop
From that our fears stopped.
Api yoga mèe kuwa idima weda kaa mèe se me-g-ea.
woman child man dem.pl all fear obl people neg come-imm-3pl
Everyone--women, children, men--from fear weren’t coming.
Nakaa-yoo ko uduma pe-seo-me-g-ea,
proximal-place pod all go-depart-if-imm-3pl
Here, everyone left,
inii nakaa-yoo ee epoo bagee yuwa too ome-gee.
1pl proximal-place house owner people det.pl only sleep-rep
only we home owners stayed on.
Ega ena naagoo koma mi-g-i.
count one day helicopter come-imm-3m
One day a helicopter came.
Ese-ani-g-i ko TJ ese-ani-g-i. Ese-ani-g-i ko
descend-unc-imm-3m pod TJ descend-unc-imm-3m descend-unc-imm-3m pod
When it landed, TJ got out. When he got out,
mee mèe weda kaa uwo kaa ese-anee see egaa wiya.
dem.sg person fear in water obl descend-unc:pla such count two
that man in fear went underwater twice.
Uwo kaa ese-ani-g-i kaboo, kaama too weda kaa bedaa mee-maa-g-i,
water obl descend-unc-imm-3m time then only fear obl again come-comp-imm-3m
When he went underwater, he then came back up,
ki-see ukuwa-g-i gaa-yaa naaki yegee-k-i, mee mèe.
ir-manner do-imm-3m be-hab adv cry-ppla-3m dem.sg man
(and) when he did like that he cried, that man.
TJ me weda kaa “Po-o-kee yoka dimi se nagi-ma-a” asi-g-i.
TJ ag fear obl go-irr-cust reason mind neg astonish-if-irr say-imm-3m
TJ in fear said, “We’ll go, so don’t worry.”
Pe-seo-mi-g-i kaboo mee mèe doo-doo-ta wa me M. doo-doo-ta pee-g-ea.
go-depart-if-imm-3m time dem.sg man carry-dep-seq 3sg poss M. carry-dep-seq go-imm-3pl
When (the helicopter) left, they took that man to his place in M.
Inii apa yuwa gembala waa ma mee-maa-ta
1pl rest det.pl shepherd 3sg and come-comp-seq
After the pastor and the rest of us came,
nakaa-yoo ome-gee too-gee, eto to-p-e,
proximal-place sleep-rep stay-rep now stay-rpst-1pl.exc
we lived on here until now.
see seto menaa mee too inii weda ke-p-e, kaama taa.
such account speech dem.sg only 1pl fear aux.be-rpst-1-pl.exc from not
That is the story of us being afraid. The end.


The following abbreviations are used in this paper.

3f = 3rd person singular feminine
3m = 3rd person singular masculine
3n = 3rd person singular neutral
adv = adverbial
ag = agent (or instrument)
approx = approximate
aux.be = auxilliary ‘be’
aux.do = auxiliary ‘do’
caus = causative
comp = completive
conjec = conjecture
cust = customary
d = dual person
dat = dative
dec = decisive
dem.pl = plural demonstrative
dem.sg = singular demonstrative
dep = departure
det.pl = plural determiner
det.sg = singular determiner
dist = distributive
dur = durative
emph = emphatic marker
exh = exhaustive
exc = exclusive
fpst = far past
gdr = go-do-return
grn = grounds
hab = habitual
if = inflectional formative
imm = immediate
inc = increment
imp = imperative
immin = imminent
indef = indefinite
inf = infinitive
ipfv = imperfective
ir = immediate relevance
irr = irrealis
iter = iterative
kin.pl = plural kinship
kin.sg = singular kinship
nfut = near future
neg = negation
np = noun phrase
obj = object marker
obl = oblique
pfv = perfective
pl = plural
pla = plural action
pod = point of departure
poss = possession
ppla = present plural action
punc = punctual
purp = purpose
rec = reciprocal
red = reduplication
rep = repetitive
rep2 = repetitive (while in motion)
rpst = recent past
rpst2 = recent past 2 (past perfect)
seq = sequential
sg = singular person
sim = simultaneous
sta = stative
term = terminate
tran = transformative
unc = uncontrolled
ynq = yes-no question marker

Auye is a Papuan language (Trans-New Guinea, Wissel Lakes) spoken by approximately 500 people living in the western highlands area of New Guinea in the Indonesian province of Papua, just north of the Me (also known as Kapauku, Ekari, Ekagi) and Wodani language groups and due east of Nabire at the southernmost part of Cenderawasih (Geelvink) Bay.

Auye has a basic 5-vowel system: a, e, i, o, u. Voiced stops /b,d/ are implosive utterance initial, and voiceless stops /p,t,k/ are slightly aspirated. The voiced velar stop /g/ is laterally released and implosive preceding front vowels /i,e/. The voiced alveolar stop /t/ also has the allophone [s] preceding front vowels. Also included are nasals /m,n/ and approximants /w,j/ (the latter written as y in the paper to conform to the orthography).

The basic syllable pattern is (C)V(V)(V). Syllables are either short or long, each having variant tone patterns. Short syllables are distinguished from long syllables strictly on the basis of timing. A short syllable has one unit (mora) of timing and a long syllable has two.

The domain of accent is the syllable as a whole. This means that even for long syllables (two moras of timing) the accent is born by the entire unit rather than just one or the other of the vowel segments. In this paper accent is only specified for certain contexts that distinguish similar tense-aspect forms (e.g. -g immediate and -'g recent past 2).

Auye has 2 emic tones (high and low) and would probably be classified as a reduced word-tone system (contrasting on only one syllable per word) according to Donohue (1997). Besides indicating the placement of the accented syllable, the lexicon also identifies the tone assigned to the word as a whole. The assigned tone (high or low) is realized at the beginning of the word (or noun phrase) and is sustained over each succeeding syllable until the accented one is reached, after which the pitch falls gradually from high to low word-finally. In this paper tone is only indicated on the low tone word mèe ‘person, man’ to distinguish it from the high tone word mee dem.sg.

I would like to thank the thirteen Auye storytellers who provided the oral and written data for this paper. I would also like to express appreciation to the community leaders for their consent in allowing the publishing of the data contained in this paper. Names in the data have been abbreviated and some of the texts in Appendix D have been shortened for the sake of personal privacy or sensitivity. I also gratefully acknowledge comments by Dr. Stephen H. Levinsohn, Dr. Rick Nivens, Dr. René van den Berg, and a blind referee on earlier drafts of this paper.

Some of the material in this paper has been documented in Moxness (2003). The author can be reached by email at mike_moxness@wycliffe.net.


The verb aspect -taa imperfective (usually glossed durative in non-dominant contexts) is actually one of six aspect 2 suffixes that specify plural action in Auye narrative. It functions as the prototypical, non-perfective tense-aspect in mainline verbs in dominant contexts, and represents the counterpart to the mainline tense-aspect marker -g immediate for non-dominant referents, as shown in (i).

...ekea me ee yiba pakaa-see-me-g-ea, anii ni me ee yiba pakaa-see-me-taa.
...3d poss house inside ascent-enter-if-imm-3pl 1sg 1sg.poss poss house inside ascend-enter-if-ipfv
‘...they entered their house and I entered mine.’

Tense and subject person-number marking is dropped, as occurs following all verb aspects for forms referencing dominant entities. Though technically it belongs to the larger group of forms marked -Ø, because of its more generic, imperfective aspectual nature, as well as its wider function as counterpart to the non-dominant mainline form, it has been separated out for special consideration in this paper. Further discussion can be found under appendices B and C.


The table in (i) displays the co-occurrence patterns of first position aspect markers with -taa imperfective and -yu perfective in second position. The numbers in the far left column signify relative perfectivity, with 1 the highest and 7 the lowest.

Distribution patterns of first position and second position verb aspect markers
Second position verb aspect suffixes
First position verb aspect suffixes-yu pfv-taa ipfv
1-asi punctualnono
2-Vpi terminal (+/- -asi punctual)nono
3-taa/-taga exhaustiveyesno
4-maa/-maga completiveyesno
6-ani uncontrollednoyes
7aspect 2 (plural action) + aux.do*noyes**
* There is a set of six plural action aspect suffixes (including -taa ipfv/dur; see note [2]), plus reduplication, that indicate plural action. The auxiliary ta/se ‘do’ is obligatory preceding any inflections, including verb aspect. See Appendix A for an overview of Auye tense and aspect.
** The auxiliary ta (se in irrealis contexts) ‘do’ actually drops out in this context, and the similarity in form may serve as evidence for a close historical link between the two morphemes. See Appendix C for examples and further discussion.

The phenomenon is known as “pluractionality” in the literature. In Auye this is expressed by various aspectual forms (a list of the most common ones can be found here in Appendix A), whereby the action of the verb is performed in some plural (i.e., imperfective) manner, and in a variety of possible contexts, including: either a singular (51b) or plural (50a) subject acting intransitively, a singular agent acting in regard to a singular patient (50b), or a plural agent acting in regard to a plural patient (compare Enemy:1 with Enemy:17, which has the same verb stem but perfective action, plural agent and singular patient).


The possibility of an underlying switch reference system was investigated, but several significant factors lead us away from this direction. First, the pertinent morphological markings are all on the final verbs, not medial ones. Second, spans of text reveal the consistent marking of same subject for both dominant referents Unruly:2-6 and non-dominant ones Unruly:18-20, while at the same time showing no variant (inconsistent) patterns in terms of dominance/non-dominance for different subject Unruly:10-17.


There happens to be another form that has the appearance of a temporal adverbial clause, but has been judged to be closer to a “topicalized” clause, and therefore has not been considered in this study. The form includes the highly perfective aspect -asi punctual (which can be phonologically separate from the verb, as in the example here) within a finite clause that is followed by the singular demonstrative mee. The “topic” marker ko pod optionally follows. The form almost always occurs at a point of high tension, such as discourse peak. The aspect -asi denotes an action that only occurs once, lending itself more readily to the process of topicalization. The high perfectivity may also be a reason why it only occurs with finite marking, even for narrator (i).

Mawai kapa-taa-(ta) eke ase-g-a mee ko yupigapi kiya...
machete chop-exh-(seq) leave punc-imm-1sg dem.sg pod middle but...
‘My chop with the machete, (it struck) the middle, but...’

At this point we would want to look into whether factors such as the degree of “agentivity” might play a role in the marking of subject in terms of dominance, considering the possibility that Numskull lacks these features. An investigation into the verbs of the Numskull text reveals that the degree of agentivity in regard to the actions of Numskull (ia) is no different than those of dominant participants in 1st person narrative (ib). (Identical verb stems are underlined.)

Amo-ki-yoo ekega-maa-g-i, noo-pegee, nu-g-i.
yonder-ir-place open-comp-imm-3m eat-rep2 eat-imm-3m
‘Nearby he opened up [the leaves that the food was baked in, then] going along and eating, he ate (it).’
Yaa ekega-maa-yu, no-taa.
wild.pig open-comp-pfv eat-ipfv
‘(We) opened up [the leaves that] the wild pig [was baked in, then] ate (it).’

The author of one oral text in the corpus was a teenager who, during the telling of his story, was being coached by others in regard to the proper wording. (The fact that the attempts were rebuffed is a separate matter.)


The subjects and agents of the actions in procedural texts are typically not expressed. The implied “dominant” participant is a generic individual or group of individuals in the community.


It is possible, however, to have habitual marking on both verbs:

...ebo yaato wa me miya mèe muni-yaa.
...big side 3sg ag also 1pl care.for-hab
‘...when [he] is big, he also cares for us.’
Donohue, Mark. 1997. Tone systems in New Guinea. Linguistic Typology 1:347‒386.
Dooley, Robert A., and Stephen H. Levinsohn. 2001. Analyzing discourse: A manual of basic concepts. Dallas: SIL International.
Givón, Talmy, ed. 1983. Topic continuity in discourse. Amsterdam: John Benjamins.
Grimes, Joseph E., ed. 1978. Papers on Discourse: SIL Publication No. 51. Dallas: Summer Institute of Linguistics.
Martin, David L. 1986. Dominance and non-dominance in Sikaritai discourse. Pacific Linguistics A 74:205‒231.
Moxness, Michael E. 2003. Auye grammar. Manuscript.
Runge, Steven Edward. 2006. A discourse-functional description of participant reference in Biblical Hebrew narrative. Ph.D. dissertation. South Africa. University of Stellenbosch.
Sheldon, Deidre. 1986. Topical and non-topical participants in Galela narrative discourse. Pacific Linguistics A 74:233‒245.