i 
He surmised that the referent for the child's first person pronouns was neither the physical 
object nor the child’s body, but the aroused self-feeling, which would indeed be the same 
referent for “I”, “me” and “mine” when used by different speakers. 
| imagine, then, that as a rule the child associates “I” and “me” at first only with those ideas 
regarding which his appropriative feeling is aroused and defined by opposition. He 
appropriates his nose, his eye, or foot in very much the same was as a plaything -by 
antithesis to other noses, eyes, and feet which he cannot control. It is not uncommon to 
tease little children by proposing to take away one of these organs, and they behave 
precisely as if the “mine” threatened were a separate object -which it might be for all they 
know. And as | have suggested, even in adult life, “I”, “me”, and “mine” are applied with a 
strong sense of their meaning only to things distinguished as peculiar to us by some sort 
of opposition or contrast. They always imply social life and relation to other persons. That 
which is most distinctively mine is very private, it is true, but it is that part of the private 
which | am cherishing in antithesis to the rest of the world, not the separate but the special. 
(Cooley, 1902, p. 162) 
These early observations and explanations by James and Cooley have been substantiated 
by subsequent research in child language development. Bain (1936) replicated Cooley's 
method and conclusions. McCarthy (1930), with refinement and replication by Goodenough 
(1938), followed Piaget's theories on egocentrism in children and reported that preschool 
children used first-person possessive pronouns far more frequently in free-play groups with 
other children than in controlled situations with adults. Schachter et al. (1974) analyzed the 
speech of 170 preschoolers in free-play settings and found claims of possession to increase 
with age, which the authors linked to improved self-other differentiation. Edwards (1978) made 
detailed analyses of scripts of child-mother interactions to conclude that chiidren use 
possession to mean differential privileges of access and use of objects. Charney (1980) and 
Greenfield (1982) both reported that the earliest and most salient usage of self-referent 
pronouns is to make claims of possession, typically of objects that do not belong to the child. 
Deutsch and Budwig (1983) reanalyzed Brown's (1973) child language data to show that 
two-year-olds use pronouns (“my”, “mine”) when claiming objects and use proper names when 
descriptively indicating ownership. Golinkoff and Markessini (1980) found that toddlers most 
easily understood possessive expressions when applied to parts of the person (e.g. “Mommy’s 
face”) and to alienable objects (e.g. “girl's shoe”). Reciprocal possession of people (e.g. 
“baby’s mommy”, “mommy’s baby") was less well understood. possibly because control and 
exclusivity are not clear in those relationships. 
In a more explicit study of possessions, dominance and the emergence of self, Levine 
(1983) measured the degree of self-definition of two-year-old boys by mirror self-recognition, 
pronoun recognition, perceptual role-taking, and pronoun production in order to pair boys of