Pitjantjatjara children’s comprehension and problem-solving in a spatial cognition task

In recent years, study of spatial language and cognition has flourished, and includes a great typological range including several Australian languages (Levinson & Wilkins, 2006; Palmer et al., 2022 inter alia). This work has predominantly had descriptive and typological aims (i.e. how do diverse languages encode spatial relations?). The Man and Tree director-matcher task (M&T; Terrill & Burenhult, 2008) is commonly used, however, due to this typological focus, typically only the director’s productions are considered. In this study, we take a novel approach to this paradigm and treat M&T as an interactional problem-solving task, not only as an elicitation tool (see e.g. Pederson, 1995 for an early precursor to our approach). In particular, we analyse children’s behaviour as matchers in adult-child director-matcher pairs. Our focus is on how children make use of the information available to them, and what this tells us about their cognitive, semantic and pragmatic development. We recorded 11 pairs of 8 adults (all F, age 30s-70s) and 11 children (8 F, age 8-12), speaking Pitjantjatjara (Western Desert; Pama-Nyungan) at Utju/Areyonga (Northern Territory); attempts to record children in the director role were not successful. We also recorded several adult-adult pairs which we do not analyse in detail here, but which showed relative uniformity across individuals’ productions. We used the OzSpace protocol (Ennever et al., forthcoming), with a reduced set of items as needed to suit children’s various developmental stages. The study was conducted as part of a broader project developing an early primary mathematics curriculum in Pitjantjatjara in collaboration with Areyonga School and community.
The adult speakers used several linguistic strategies to describe spatial relations. The predominant strategies are shown in Table 1 with their possible frames of reference (FoRs); these combine with a rich case-marking system. Other strategies were present but infrequent. We analysed how speakers refer to orientation vs location, but also whether the description (considered cumulatively) completely disambiguates the target card from others in the set. Examples (1-2) show a ‘complete’ and an ‘incomplete’ description respectively (with speakers facing west). There is an ambiguity in (2) due to the allative -kutu, which can indicate orientation or a vague location (Goddard, 1985, pp. 43–44); here the speaker intends it to refer to orientation only.
As well as the child matchers’ speech, we analysed their selection of cards and whether this matched the information provided thus far by the director. We also analysed all intentional touches of other cards as the children considered which one to select. An illustrative example is shown in extract (3). In line 1, the speaker produces an incomplete description; kuranyu ‘ahead’ and maḻa ‘behind’ can have either relative or intrinsic FoRs. In this case, the description is true for both FoRs, but there are other cards for which only one or the other FoR would also license a valid interpretation, hence at this point neither orientation nor location can be known by the matcher with certainty. In line 2, this ambiguity is resolved with ngalya- ‘hither’ indicating orientation. In line 3, the child (F, 12;3) touches a card which has the correct location, but the opposite orientation to the target card. This matches a relative interpretation of line 1, and while the orientation does not match the description in line 2, it is notably on the correct axis. In line 4, the child selects the correct card, having pieced together the information from all three clauses. Detailed coding and analysis of all 147 trials provides insight into errors, as well as the different types of logical and pragmatic inferences made by children. This in turn gives a preliminary picture of Pitjantjatjara children’s developing spatial language at ages 8-12.
As well as being the first dedicated study of spatial language in Pitjantjatjara, we make a number of novel contributions, including considering the M&T task as an interactional problem-solving task (from the perspective of game-theoretic pragmatics; Benz et al., 2005), and adding to the very limited research on the development of children’s spatial language and cognition in non-WEIRD settings.