The development of support intuitions and object causality in juvenile Eurasian jays (Garrulus glandarius)

Abstract

Knowledge about the causal relationship between objects has been studied extensively in human infants, and more recently in adult animals using differential looking time experiments. How knowledge about object support develops in non-human animals has yet to be explored. Here, we studied the ontogeny of support relations in Eurasian jays (Garrulus glandarius), a bird species known for its sophisticated cognitive abilities. Using an expectancy violation paradigm, we measured looking time responses to possible and impossible video and image stimuli. We also controlled for experience with different support types to determine whether the emergence of support intuitions is dependent upon specific interactions with objects, or if reasoning develops independently. At age 9 months, birds looked more at a tool moving a piece of cheese that was not in contact than one that was in direct contact. By the age of 6 months, birds that had not experienced string as a support to hold up objects looked more at impossible images with string hanging from below (unsupported), rather than above (supported). The development of support intuitions may be independent of direct experience with specific support, or knowledge gained from interactions with other objects may be generalised across contexts.

Figure 1. Still frames of the video stimuli.

The arrow indicates direction of movement from the starting position. Birds received six video stimuli: contact (the cheese was in contact and in front of the direction of the moving rake, incorrect contact (in contact and behind the direction of the moving rake, and no contact (not in contact with the rake, but moving alongside the rake. For each of the three categories of support, the cheese was either positioned outside or inside of the rake. Therefore the position of the cheese relative to the rake differed for each contact category. Birds were first presented with a familiarisation video before the test stimuli to show the rake moving independently from the cheese (see methods and supplementary materials for videos).

Figure 2. Static images of support presented to birds at age 6 months.

(a) correct support (supported in the centre of the box), (b,c) insufficient support (a greater surface area of the cheese was off the box than on the box), (d,e) no support (the cheese had no contact with the box and was suspended in mid-air) and (f,g) incorrect support (the surface area of one side of the cheese was in complete contact with the vertical edge of the box). Each set of support categories is duplicated with either a string up (could be supported by string, a,b,d,f) or a string down (not supported by string, c,e,g). Birds at age 3 months were presented with four images of the same four categories of support, but without any string (see methods).

Figure 3. Looking time results for videos stimuli across all age stages (3, 4, 6 and 9 months).

Cheese was positioned a) outside of the rake, GLMM age*no support Est. 1.03 ± 0.36, z = 2.84, p = 0.005; age*incorrect support; age 9 months Est. −1.00 ± 0.25, z = −3.64, p = 0.0003 and b) cheese was positioned inside the rake. GLMM age 4 months Est. −0.33 ± 0.13 SE, z = −2.57, p = 0.01; 6 months Est. −0.23 ± 0.12 SE, z = −1.94, p = 0.05; 9 months Est. −0.93 ± 0.03 SE, z = −6.44, p < 0.0001. Contact and age 3 months were set as the reference categories. Whiskers denote 1 standard error.

Figure 4. Looking times at age 3 months.

GLMM insufficient support Est. 0.12 ± 0.2, z = 0.61, p = 0.54; no support Est. −0.11 ± 0.2, z = −0.57, p = 0.57; incorrect support Est. −0.29 ± 0.06, z = 0.22, p = 0.83). Contact (control) was set as the reference category. Whiskers denote 1 standard error.

Figure 5. Looking times at age 6 months with string up (SU) and string down (SD) images.

(a) nail group: GLMM no support SD (Est. 0.56 ± 0.28, z = 2.00, p = 0.045) and insufficient support SD (Est. 0.53 ± 0.26, z = 2.03, p = 0.043); (b) string group: GLMM no contact SU (Est. 0.56 ± 0.28, z = 1.99, p = 0.046) and insufficient support SU (Est. 0.53 ± 0.28, z = 1.89, p = 0.058). Contact (control) was set as the reference category. Whiskers denote 1 standard error.

Figure 6. Familiarisation stimuli presented before each session.

(a) videos, (b) core knowledge images, (c) string support images.