Young infants' awareness of object identity and number: Evaluating arithmetic reasoning, object file, and object tracking accounts.

Wynn (1992) made the controversial claim that young infants understand addition and subtraction, finding that they looked longer at events with numerically incorrect outcomes. Although alternative interpretations of Wynn's influential results have been offered, investigations have stopped short of the systematic manipulation of relevant factors needed to identify an appropriate interpretation. This proposal is motivated by the pressing need to provide a firmly based account of how young infants process single and multiple object occlusion events.

Alternative explanations include the object file account, which although based on representational processes does not assume a symbolic understanding of number. It does, however, make predictions regarding the outcome of certain manipulations of Wynn's task, but these predictions conflict with each other. Secondly, a lower level object tracking account makes specific predictions regarding infants' responses to violations of existence, featural identity, and position of objects in Wynn's task, and is parsimonious through not assuming mental representation.

It is apparent that manipulation of Wynn's task has potential to answer key questions regarding infants' object tracking that go well beyond Wynn's initial question. This proposal involves a core set of 23 experiments with 5-month-olds, as a systematic investigation of infants' ability to keep track of objects. Our main measure will be infants' looking time as an indicator of violation of expectation. (The assumption of this well tried measure is that infants look longer at events that violate their expectations.) This will be supplemented by eye-tracking and pupillometry data to investigate whether infants look more at an object that violates expectation regarding its existence, featural identity, or position. In cases where young infants do not respond to a particular violation, the experiments will also be run with a group of older infants to seek evidence for developmental increase in sensitivity.

Experimental series 1 and 2 will involve a single object, and will involve the numerical operations 0 + 1 = 1 and 1 - 1 = 0, respectively (see Annexe 2 for a figural description). The 0 + 1 experiments will begin with an empty stage whereupon a screen will rotate to vertical, and an object will be introduced and placed behind the screen. On violation trials the screen will rotate down to reveal, (i) no object, (ii) the object in the wrong place, (iii) a different object in the right place, or (iv) a different object in the wrong place. The 1 - 1 Experiments will commence with an object present, the screen will rotate to hide it whereupon a hand will appear and remove the object. On violation trials, the screen will rotate down to reveal one of the same sorts of violation as in the 0 + 1 case. Series 3 and 4 relate directly to Wynn's work, involving two objects and concerning the numerical operations 1 + 1 = 2 (initial state one object present, screen rotates whereupon a second object is added) and 2 - 1 = 1 (initial state two objects present, screen is rotated whereupon one object is removed). In the case of series 3 and 4, we shall assess the degree to which infants detect different forms of violation applied to the first versus the second placed objects. In the 1 + 1 case we shall also assess performance on a variant in which both objects are added once the screen is rotated, to test the object file account prediction that an object perceived in position before occlusion leads to a stronger object file representation than one seen disappearing behind a screen.

We anticipate that these studies will provide an important body of evidence arising from systematic variations of a single procedure. This should clarify apparent contradictions in the literature and will provide a superior account of the processes through which young infants keep track of multiple objects.

Although this is basic research, the work is likely to impact positively on the knowledge base of parents of infants and of those who care for infants such as health visitors, midwives, and child minders, through providing a fuller understanding of the development of early perception and knowledge and particularly through providing insights concerning young infants' visual ability. And as a result this information is liable to have a positive impact on infants themselves. This outcome will be achieved locally in a very direct way through providing information to parents bringing their infants to our laboratory, and also through the insights provided to parents by observing our work. In particular, it is our experience that on observing eye-tracking records, parents are amazed and delighted by the accuracy with which their infant scans the visual scene and follows moving objects. This insight into their infant's ability is liable to lead them to present enhanced visual experiences better adapted to their infant's ability, something known to have a positive effect on development.

We will take steps to disseminate information about infant perceptual ability, particularly through providing examples of infants' visual tracking of scenes, to those who care for infants inside and outside the NHS. Also, more widely, but on a longer timescale, parents and carers will be exposed to new findings through the media: the findings bear directly on a major issue of public concern, the Nature-Nurture issue (e.g., as illustrated in the popular 'Child of our time' TV series and other media presentations) and will almost certainly attract media attention. Additionally, as these findings are published and appear in textbooks, students of psychology and other relevant areas such as nursing and paediatrics will benefit from the fuller understanding of infant development that these should bring. In turn, this increased knowledge will filter through to parents and the general public.

The attached document 'pathways to impact' indicates the steps we will take to maximise the impact of our activities beyond the academic sphere.