This short essay seeks to describe some contemporary cognitive perspectives of the underlying processes concerned with the development of ‘objecthood’ in infants. Here the dynamic processes responsible for the infants’ ability to mentalise the world of objects are held as freighted by the term representation (Ger. Vorstellung). In concert with the description of infants’ representation of objects, reference will be made to evidence admitting from developmental studies of the former processes seen through the lens of object individuation paradigms (i.e. peripheral-origin and core-origin). The first section traces and discusses Piaget’s notion of acquisition and study of knowledge – genetic epistemology. The second section discusses more recent deviations from the Piagetian peripheral-origin account, situating the latter as foundational to the later discussion. The third section aims to build and extend upon the core of the first and platform of the second sections to reveal a whispered call for the privileging of individual difference.
Some monumental figures of the constructivist paradigm stand tall in the hall of ones attention (e.g. Hegel 1969; Kant 1999; Darwin 1998) as one approaches Jean Piaget’s (b. 1896 – d. 1980) chosen term for his synthetico-empirical theory (see Piaget 1972) for the study of the acquisition and growth of knowledge – genetic epistemology (Fr. Épistémologie génétique). More commonplace is perhaps a notion of an elastic, intimate relation (Piaget 1952; 1954; 1973) held to be coextensive between what we may perceive and how we act upon it; and conversely, what we may act upon and how we perceive that also. From the basis of an understanding of an intimate reciprocity, Piaget’s description of the acquisition and growth of knowledge is one entirely formed from reciprocal constructs of understanding (accommodation and assimilation). Knowledge itself is pointed up as an emergent psycho-biological adaptation, a synthesis born of the gradual experiencing of the immediate environment exclusive to the combined schema of assimilation and accommodation (adaptation). This periphery inwards understanding for the acquisition of reasoned knowledge (cognition) is more technically termed the peripheral-origins thesis (e.g. Berkeley 1975; James 1982; von Helmholtz 1926). And here, one might almost, incorrectly, assume oneself back among the myths of tabula rasa (e.g. Aristotle 1988; Aquinas 1993; Locke 1997; Skinner 1959). Almost, yet, not quite, as Piaget was certainly neither for behaviourism nor nativism. For him the balance (equilibration) and organisation (schema) of perceptions and actions were held to be reasoned cognitive processes acquired by adaptation (accommodation and assimilation) to the immediate environment.
To Piaget the infants’ sense organs direct motor actions and in combination knowledge of the immediate environment is acquired. Piaget thus refers to infancy as the sensorimotor stage spanning ≤24mths. The sensorimotor stage is further divided into six sub-stages. As secondary circular reactions (i.e. repeated actions characteristic of sub-stage 3 between 4-8mths) give way to coordinated secondary circular reactions (sub-stage 4 between 8-12mths), the abilities of object and event mentalisation are suggested to replace those attained at the preceding stage. Which is to say, each stage overcomes the previous stage throughout the developmental stage theory model (i.e. sensorimotor (0-24mths) → pre-operational (2-7yrs) → concrete operational (7-11yrs) → formal operational (11/12yrs+). Mentalisation abilities, which Piaget called representation, reach fruition between 18-24mths in the forms of symbolic representation and inchoate semantic representation. Piaget’s descriptions have been replicated widely across cultures (see Dasen 1973) and became an established standard by which child development has been assessed.
Robust challenges to Piaget’s constructivist theory and methods however, from various different laboratories point to: neonates imitating the facial dispositions of an adult (see Meltzoff & Moore 1977); infants as young as 3.5mths perceive the immediate environment as a consistent phenomenal plane of objects; and, infants indeed acting on this plane adaptively, however, with maturation capacities extended and refined – not merely overcome. The following discussion obtains from such findings.
Deviation from Piaget: Violation-of-expectation method
Researchers have suggested influential findings indicating infants face numerous difficulties when searching for hidden objects which serve illustrate to what is and what is not understood (e.g. Bower 1982; Bower and Wishart 1972; Diamond 1991; 2000).
Bower (1982) has demonstrated that infants as young as three or four months showed surprise when a once visible, then occluded, object had seemingly ‘disappeared’ after only a few seconds. Immediately the suggestion was made that object permanence was present much earlier than predicted by Piaget’s sensorimotor sub-stages allowed for. Further, Diamond (1991; 2000) has utilized temporal variations between the time before switching an objects location (A to B) and the delay before the infant could reach for an object. Her findings demonstrate that immediate ‘reach’ opportunities for 7.5mth resulted in correctly identified object location. Whereas, 7.5mth could not withstand a ≤2 sec delay without incorrect location choice. 9mth infants were found to withstand a ≤5 sec delay prior to successful object location, likewise, 12mth infants a ≤10 sec delay. Once again therefore infants were demonstrated as capable of representation of occluded objects; though the infants’ working memory is demonstrably limited. It is interesting that Diamond noted motor habit (perseveration) played a significant role in the obfuscation of knowledge of the hidden object during incorrect location choice – where the eyes focused attention (B) was observed as inconsistent with where the arm and hand reached out (A). These motor habit findings were revisited and modified by Thelen et al. (2001). In this experiment no objects were hidden. Rather, an object at location A is placed within reach of the participating infants. After habituation to the first object a second object was placed at location B. Infants continued to reach toward the ‘old’ object (location A) whilst their attention was drawn to the new object (location B). Thelen et al. (2001) provides compatible support for motor habit data in support of Diamond’s (1991; 2000) findings.
Others reasoned differently (e.g. Baillargeon 1998; Spelke et al. 1992; Baillargeon et al. 1985), deciding that memory and motor habit might be circumvented altogether by utilizing infants’ well-documented tendency to stare at novel stimuli. Baillargeon et al. (1985) devised the 180o hinged-screen differential looking experiments where infants’ expectations were violated by impossible events (i.e. the hinged screen seemingly passing through a yellow box): and in so doing created an influential and revealing new paradigm for testing object permanence by way of impossible events. That is, impossible spatiotemporal events which undermined natural physical laws and, seemingly attracted a rent in the infants’ sensibilities leaving them noticeably surprised.
Of these early studies using differential-looking as the dependent variable a seminal paper by Spelke et al. (1992) is interesting and noteworthy. The principle findings garnered from six habituation trials and six control test trials indicate a non-significant looking time in experiment 4, and a significant mean looking time in experiment 5 (with those of experiment 1 (Wilcoxon-Mann-Whitney z = 2.26), experiment 2 (WMW z= 3.88) and experiment 3 (WMW z = 3.00)). It is noteworthy perhaps that ANOVA for expts. 4 or 5 did not produce statistically significant results.
The researchers posited that their own and other similar findings (e.g. Baillargeon 1990; Kim & Spelke 1992; Spelke et al 1992) were compatible with two principles: inferences about object motion; and, inferences about object continuity and solidity are not dissimilar to those seen in common adult conceptions. The peripheral-origins thesis had been confronted with some problematic but productive sources for questioning: Might the basis of cognition develop from an innate core rather than from perception and action? Moreover, did initial (core) cognitive abilities yield appropriate conceptions that are relevant to immediate experience; that is, was it not plausible that during maturation, as knowledge grows, that conceptions might enrich and refine initial core abilities? Spelke et al. (1992) core-origins thesis provides a persuasive alternative to the peripheral-origins thesis.
None the less, some researchers (Cason & Cohen 2000) maintain that results yielding from studies using differential-looking may have misattributed cognitive ability to young infants. In a recreation of a ‘classic’ study (Baillargeon et al. 1998; Baillargeon 1985) using the 180
hinged-screen differential looking experiment with 8mth infants the researchers found that infants could be dishabituated to stare longer at novel events whether possible or impossible ((experiment 1) familiar angle novel block ≤8 secs, (experiment 4) novel angle and block ≤9 secs).
Taken as a whole, the above brief discussion opens a small aperture through which some results are demonstrably contrary to those predicted by the landmark cognitive research of Piaget (1954). Subsequent deviations from peripheral-origin theory are also seen to yield results of some interest to contemporary cognitive development researchers made possible by removing means-end coordination tasks (e.g. the differential looking paradigm using the violation-of-expectation method).
Intermodal perception of object individuation in 10mth old infants
The modest discussion of the representation of objects above has, thus far, been implicitly indicative of 3-8mth infants’ abilities rather than explicitly indicative of infants’ cognitive abilities at around 10mth (though the empirical data has been consistently explicit throughout). The remaining discussion seeks to build upon the foundation of the former section. Earlier the suggestion was tendered that infants’ understandings obtaining from the individuation of objects were influenced in part by spatiotemporal continuity (see Spelke 1995; Spelke et al. 1992; Xu & Carey 1996). Some researchers (Káldy & Leslie 2003; Leslie et al. 1998) have estimated these findings to be central to the suggestion that infants’ object cognition might be reliant on an object indexing system. In other words, researchers (Káldy & Leslie 2003; Leslie et al. 1998) have espoused that spatiotemporal information (either by location or feature) binds an object to an index, given that an index has been assigned for such object bindings to take place (e.g. by shape or colour). Put rather crudely, it might be called a cascade activation model with limited categorical indexes available. Here an object (O
) activates an index (I
) which remains activated for a period dependent on the infants’ object-based attention threshold. The onset of a second object (O
) activates a second index (I
). In this example, the appearance of the second object might be met with by an expectant infant seeking to match these two objects (O
) with two activated indexes (I
). Compatible findings (Xu & Carey 1996) were found for this thesis put forward by Leslie et al. (1998) and Káldy & Leslie (2003). In Xu & Carey’s (1996) studies (N=32, x16 10mth, x16 12mth mixed-gender sample all from Greater Boston) they consistently found 10mth infants’ were unable to establish representations of two numerically distinct objects using property/kind information as a guide to success. In the spatial condition (two at a time or alternately) and in the temporal condition (one at a time) objects are shown and replaced from behind a screen by researchers. The two objects shown are always a different ‘kind’ from each other (e.g. cup and shoe). For the 10mth babies there was as a significant time difference for two-object (M=9.8 sec) over one-object displays (6.7 sec). In the 12mth babies there was a marginally significant time difference for two-object (M=9.6 sec) over one-object displays (8.4 sec). However in the 12mth babies there was also a significant interaction (at 95% confidence) between trial type and outcome (F(1,15)= 4.824, p=0.05). These findings apparently indicate that the two groups established individuation of two distinct objects using different capacities of kind/property. That is, 10mth babies used location alone to individuate objects whereas 12mth babies relied on spatial and temporal conditions. It is the finding that familiar sortals (e.g. duck, truck, cup, ball, book and bottle) do not provide sufficient criteria for object individuation which is perhaps quite interesting and requiring further research in this age range (10-12mths).
Convergent findings from object individuation/identification research (Van de Walle et al. 2000) dovetails well with differential looking time data (Xu & Carey 1996) for infants’ object individuation (at 10mths) and identification (at 12mths). Here individuation is considered as freighting detection of an object and the representation of that object, whilst, identification freights with it the further step of taking object representation and binding it to object identity (i.e which one?) following the convention used in adult attention literature (see Leslie et al. 1998). Researchers employed for the first time the manual-search paradigm; favouring this method for its relative directness, and its ‘promise’ as a new bases from which to study object individuation (Van de Walle et al. 2000).
In the first experiment Van de Walle et al. found 9mth infants were, as predicted (Xu & Carey 1996), seen to be significant in respect of age and outcome type in their ‘reaching’ for objects (N=64, random sample population). 46 of 64 infants reached frequently or more on two-object/unexpected outcomes, and 49/64 infants reached frequently or more on two-object/unexpected than two-object-expected outcomes (Wilcoxon z = 3.90 and 4.12 respectively) at 99% confidence (p0.001).
In the second experiment Van de Walle et al. 12mth infants (N=16) reached more purposefully when only one of two possible objects had been retrieved. 10 of 16 infants reached more often on two-object/unexpected than one-object/expected outcomes (Wilcoxon z = 2.2) at 95% confidence. The one-represented-object hypothesis was seen to be unsustainable in this test. Experiment two might be said to be compatible with the suggestion that 12mth infants’ represent precisely the number of hidden objects in cases of either one or two.
Another quite interesting study (Feigenson et al. 2002) also utilized a variation of the manual-reach paradigm to some notable effect. In their study of 10 and 12mth infants Feigenson et al. focus on the underlying numerical analogue-magnitude representations in choices of more; specifically in this case, more crackers of various sizes placed sequentially to ascertain the presence or not of set-size signature of object-file representations (expt1. N=124, 50/50 split of 10 and 12mth; expt 2. N=48, 18/30 split; experiment 3. N=32, 5/27 split). The results from experiment 1 showed infants from both groups choosing more over less in the 1-versus-2 condition (80% or 13/16). However in larger number conditions (e.g. 3-versus-4) there was no significant difference in ability (50% and 40% respectively). The results from experiment 2a explored failures in experiment 1 owing to motivation reasons. Results for 2a showed that infants from both groups chose more over less in the 3-versus-6 condition (80% or 13/16). This finding suggests that lack of motivation was not a factor. The results from experiment 2b explored failures in experiment 1 owing to complexity. Results for 2b showed that infants from both groups chose more over less in the 1-versus-2 condition when the stimuli received twice the usual amount of motion and presentation complexity (75% or 12/16). This finding suggests that complexity was not a factor. The results from experiment 3 explored whether choices in experiment 1 showed infants from both groups choosing more over less in the 1-versus-2 condition due to number of items or total amount of cracker. Expt 3a was a greater-area condition and experiment 3b an equal-area condition. Results for experiment 3a showed infants choice based on quantity (a huge cracker) and not number of items (75% or 12/16) Results for expt 3b showed infants choice based on equal-are were at chance (37.5% or 6/16). The overall pattern of both successes and failures were suggested to compliment and help make sense of the object-file hypothesis proposed by others (e.g. Káldy & Leslie 2003; Leslie et al. 1998; Uller et al. 1999).
Reflections in conclusion
In seeking to satisfy the stated purpose (attending to infants’ representations of objects) concerning object individuation and identification one ought to perhaps remain mindful that the unique richness of individual difference is neither diminished nor lost sight of through reflective and responsible handling of information obtaining from the unfolding tapestry that is human patterning.
- John Tropea: The Way We Were: 10 Crucial Child Psychology Studies – PsyBlog (spring.org.uk)
- Jean Piaget’s theories (slideshare.net)
- Developmental Psychology G1 (slideshare.net)
- Jean Piaget – Psychologist of the Week (psychology.about.com)
- Piaget’S Theory Of Cognitive Development (slideshare.net)
- What Are The Origins of Number Representation? [The Thoughtful Animal] (scienceblogs.com)
- Importance of Early Childhood Care and Education (socyberty.com)
- Components of Human Development (slideshare.net)