Object naming induces viewpoint-independence in longer term visual remembering: Evidence from a simple object drawing task

Abstract

The impact of object naming on object drawing confirms an association between object categorisation and viewpoint-independence in longer term visual remembering. Adult participants viewed a novel object from a viewpoint from which it would not normally be drawn from memory. The experimenter either labelled the object with a novel count noun (“Look at this dax”) or did not (“Look at this object”). Participants then drew the object from immediate, short-term, or longer term memory, with no constraints being imposed on how they should depict the object. When the object was named at presentation, but not otherwise, the transition from immediate to longer term remembering increased the likelihood that the object was depicted from a viewpoint from which it had not been seen. This trend was reversed when participants were asked to depict the object in the orientation in which it had appeared to them. These results are discussed in relation to an account of the conditions under which visual category representations become established and may be used preferentially over image-like visual representations.

Notes

¹As a referee has correctly pointed out, a GSD could have an associative link to a representation of the typical colour for a category of object, and evidence from neuropsychology indicates this to be the case. However, notwithstanding the presence of such links, structural descriptions and colour will still be separately represented, allowing stored information about structure to be accessed selectively, without information about colour being accessed at the same time. Our argument is that the same selectivity is not an option when images are accessed, because images do not represent structure and colour separately, but rather in an integrated manner. In addition, of course, many object categories, especially artefacts, do not have a typical colour, so their GSDs will not be linked to representations of colour.

²A count noun is a noun for a single entity, including objects (lamp) and mental entities (image), that can be counted. Hence, count nouns appear with indefinite articles in phrases referring to a single instance of a type of entity (a lamp), and in phrases referring to multiple instances of a type of entity (six lamps, many images).

³With regard to why, when an orientation has to be specified for a category of object, certain orientations become preferred, we presume there are many contributing factors. Frequency of exposure to particular orientations would be the most obvious, though this cannot apply here. The ease with which the constituent geons can be discriminated from a particular orientation is claimed to be another factor. In addition, as a referee pointed out, the stability of an object in a particular orientation is likely to be another factor, in part because an unstable orientation is an unlikely orientation from which to encounter the same type of object in the future. Having said this, however, we might note the overwhelming tendency for people to draw bicycles in an unstable orientation (i.e., upright, and with no apparent support being depicted). The functional significance of certain orientations also is likely to be very important. Thus, drinking mugs can only serve their usual function from a limited range of orientations, and they normally do this best when their handles are pointing to the viewer's right (so that a right-handed person can grasp the mug easily). We intend to run a study in which we explain an intended function for our novel object. In one condition, this function will require the object to be oriented with the attached part pointing down (e.g., so that indentations can be made in pastry). We then expect this orientation to become the preferred one, and the one towards which drawers will shift when object naming is combined with longer term remembering.

⁴A good illustration that object drawing can reveal the varying contributions of different types of visual representation according to memory delay is provided by Bozeat et al. (2003). They presented geometric designs and pictures of familiar objects to patients with semantic dementia. They asked the patients either to copy the pict ures, or to draw them some time after they had been removed from view. With regard to the pictures of familiar objects, the patients produced good copies, but their drawings after a delay were much less good. In addition, the quality of an individual patient's delayed drawings was correlated with the severity of their dementia, and with the quality of the drawings they produced of the same objects in response to being given their names. There was no unusual deterioration with delay in their drawings of the geometric designs. The point to note in the context of the present study is that whereas immediate object drawing was uninfluenced by category representations in long-term memory, and hence by the dementia, delayed object drawing was.

⁵It is beyond the scope of the present paper to explore either the nature of the processes required to derive each type of visual representation, or the interdependencies between them. Though they are likely to share very early visual encoding pathways, it is not clear at what point they dissociate. It is also unclear if the processes dissociate to operate in parallel of each other, as illustrated in Figure 3, or if the processes responsible for deriving a viewpoint-independent structural description follow in sequence those responsible for deriving the corresponding viewpoint-specific image(s).