Frames of spatial reference in five Australian languages

ABSTRACT

Australian Indigenous languages are widely cited as depending overwhelmingly on abstract cardinal terms for spatial reference. However, considerable under-recognized diversity exists, with systems invoking aspects of local topography or egocentric projections.

The first step toward an empirically grounded understanding of the wider implications of Australian spatial reference systems is to establish what components of spatial systems actually occur in what combinations across the continent. This article examines the spatial systems of five Australian languages to test hypotheses about the role of the environment in shaping linguistic representations of space, revealing under-recognized aspects of Australian systems, including the use of egocentric (“relative”) reference frame but only on the sagittal axis; a nearside-farside system; and multiple systems invoking diverse salient environmental features.

KEYWORDS:

• Spatial language
• Australian languages
• frames of spatial reference
• Topographic Correspondence Hypothesis
• Sociotopography

Notes

¹ We are grateful for the comments of the guest editors, three anonymous reviewers, and Jonathon Lum. We are grateful to Nick Reid and Mark Harvey for discussions on Ngan’gityemerri and Wagiman respectively, and to Claire Bowern for comments and data on Bardi. All errors remain our own. The research reported on here was supported by the Australian Research Council DP200101079, and the University of Newcastle, Faculty of Education and Arts, Center for 21^st Century Humanities.

² An exception being abstracted upriver-downriver terms noted in the Daly River region (Levinson, 2003, p. 336, note 3; Schultze-Berndt, 2006) and occasionally elsewhere (Schultze-Berndt, 2006, pp. 63, 103).

³ Morpheme-by-morpheme glosses of data in this article are standardized and slightly simplified to make the data more accessible to linguistic non-specialists. Abbreviations used largely conform to the Leipzig glossing rules and are: 1 “1^st person”; 2 “2^nd person”; 3 “3^rd person”; abl “ablative case”; all “allative case”; anaph “anaphora”; aug “augmented number” (roughly = plural); aux “auxiliary verb”; cl “noun class”; cont “continuous aspect”; dir “directional”; dist “distal”; du “dual”; excl “exclusive”; foc “focus”; frdist “far distal”; fut “future tense”; hab “habitual aspect”; imp “imperative mood”; instr “instrumental case”; ipfv “imperfective aspect”; irr “irrealis modality”; itr “detransitiver”; loc “locative case”; masc “masculine gender”; min “minimal number” (roughly = singular); neg “negative”; neut “neuter gender”; npfv “non-perfective aspect”; nrdist “near distal”; nsg “non-singular”; obj “object”; obl “oblique case”; part “participle”; past “past tense”; pfv “perfective aspect”; prf “perfect aspect”; perg “pergressive case (= ‘via’)”; pl “plural”; pres “present tense”; prox “proximate”; punc “punctual aspect”; real “realis modality”; sbjv “subjunctive mood”; sg “singular”; subj “subject”; vis “visible”.

⁴ Unfortunately, Bohnemeyer et al.’s (2015, p. 175) table summarizing frames of reference appears to place absolute outside geocentric. It is clearly intended to be a geocentric subtype (see e.g. Bohnemeyer & O’Meara, 2012, pp. 218–220).

⁵ One anonymous reviewer pointed out that literacy may be a confound in this. It would be worth considering in future research a correlation between literacy, with its alignment along the transverse axis from the perspective of the reader, and a preference for (or even availability of) a transverse axis in languages with a tradition of widespread literacy.

⁶ For example, Warrwa is described as conforming to the “typical” Australian pattern of using absolute rather than “relative, speaker-based, frames of reference”, because terms for left and right are not used to specify a location or direction with respect to a ground (McGregor, 2006, p. 148), and Majid et al. (2004, p. 112) classify Warrwa as lacking relative FoR. However, the language does employ terms for front and back in both an intrinsic and relative way (McGregor, 2006, p. 130). Similarly, speakers of Jaminjung are reported to not make use of relative FoR because terms for left and right are not used projectively (Schultze-Berndt, 2006, p. 103) although terms for front and back can have relative projective uses (Schultze-Berndt, 2006, p. 109; Hoffmann, 2019, pp. 7–8).

⁷ Local nouns in Ngan’gityemerri are defined by their ability to take ablative case (“from”) but not locative (“at”) or allative (“to”) case.

⁸ Hoddinott & Kofod (1988) represent the instrumental enclitic as =ningki, rather than Reid’s (1990) =ninggi, and the 'high' term as gangki rather than Reid’s ganggi. These are purely orthographic differences and Hoddinott & Kofod’s data has been modified here for consistency.

⁹ The status of the vertical axis in frames of reference is not crucial here. Data such as (10a-b) in this section and other sections below is given only to show that terms used for topographic elevation in geocentric FoR are those also used with the vertical axis.

¹⁰ As in Ngan’gityemerri, Wagiman local nouns are defined by their ability to take ablative but not locative/allative case.

¹¹ The proximal and far distal forms include the deictic roots mayh- and magu- that occur in other morphological constructions. However, the near distal root muny- only occurs in this nearside-farside construction. In other constructions the near distal root gayh- occurs (see, e.g. the ablative forms mayh-gunda, gayh-gunda and magu-gunda). Muny- is perhaps connected historically with munya ‘low, inside' (see below), although no elevation or containment is synchronically encoded.

¹² The data is insufficient to determine whether these are local nouns, however the lack of overt case is consistent with that status or that of adverbs.

¹³ It is not possible to determine from the data whether or not these terms are grammaticized as local nouns, however, the presence of the pergressive suffix –ba in (26) indicates they are nouns of some sort. This is not inconsistent with status a local nouns, as the pergressive in Ngan’gityemerri occurs with local nouns, even stacking with the ablative.

¹⁴ The transverse coverbs also occur as nouns. The sagittal and transverse nouns do not appear to be local nouns as they display the same case properties as other common nouns.

¹⁵ These examples are not intrinsic projections from the ground’s own left and right. In (32a), for example, the ball is in front of the chair’s intrinsic front in the stimulus photograph (Hoffmann 2013, p. 3).

¹⁶ Directions toward a location are encoded in MalakMalak with either the locative/allative case suffix -yinnga/-na(na) or the directional case suffix -nen/-(y)en. The locative/allative suffix encodes a static location or endpoint of a telic motion event. The directional case suffix encodes the direction of an atelic, ongoing or future motion event or of orientation.

¹⁷ A third, [miri(-nen) kantjuk] “zenith; midday; north” (lit. “sun(-dir) be.on.top”) is not attested outside of elicitation.

¹⁸ These terms take ablative case but are not attested with locative or allative, consistent with status as local nouns.

¹⁹ Brown (1983, pp. 132–133), however, identifies wind names as a source of what he regards as cardinal direction terms in several languages (none Australian), principally for “north” and “south”.

²⁰ We are grateful to the Australian Research Council for funding under DP200101079 supporting the OzSpace project.

Additional information

Funding

This work was supported by the Australian Research Council [DP200101079], and the University of Newcastle, Faculty of Education and Arts, Center for 21st Century Humanities.