Corinne Varon-Green, Ed. D. <varonco263@yahoo.com> |
| www.varonfinearts.com |
© 2008 varon-green All rights reserved. No parts of these pages may be reproduced or transmitted in any form or by any means.
Corinne Varon-Green, Ed. D. <varonco263@yahoo.com> |
| DISSERTATION 2004 Harvard University |
© 2004 varon
IMPRINTS OF SELF
Manifestations Of Bilingual And Monolingual Children’s
Cognitive Structures In Drawings Of Self And Significant Others
by Corinne Varon, Ed. D. Harvard University
It is my conviction that each drawing is the reflection of the personality of its maker; that it
expresses affective aspects of the personality as well as cognition… that the approach of the
examiner must of necessity be intuitive as well as analytical. (Di Leo, 1983, p. 60).
For my dissertation, I conducted a comparative study of drawings to describe how the nature of
self in interaction with others differs when produced by distinct groups of English monolingual,
Spanish monolingual, and bilingual speakers enrolled in two different elementary school programs.
Three drawing tasks were administered to all 5th and 6th grade students enrolled in one public
school in Massachusetts.
This exploratory study examines the cognitive structures underlying the construction of drawings
that capture representations of self and others. I examined the complexity of children’s cognitive
structures as represented in their graphic, visual, and verbal forms. This study also tested the
practical applications of a visual scale derived from cognitive measures based on behavioral
observations and verbal narratives. Children’s autobiographical visual and verbal recollections and
events associated with happy, sad, and angry supplied the data.
Davis (Lawrence-Lightfoot & Davis, 1997) affirms how art is the means to understand thought in
The Art and Science of Portraiture,
Through internal symbols or representations that are vehicles of thought, the producer of art
constructs a world view (p. 29).
Art allows children to reflect about their life, giving it a meaning while they communicate it to
others. Visual arts manifest ideas and feelings that may not be revealed through words or actions.
“Art puts into work, energies and contradictions that we would not otherwise be able to tap into"
(Lima, 1997, lecture).
To systematically compare performances, 92 subjects were selected. I compared up to 4 distinct
groups of students ages 10-13, grouped by their linguistic classifications and/or program
enrollment. Groups were also matched by their level of intelligence as measured by the Raven
Progressive Matrices scores (Raven, 1986).
Drawings were scored according to established ranked criteria in three categories graphic
composition, visual representation, and verbal narrative. The scales for graphic composition and
verbal narrative have been widely used to measure artistic and cognitive development. For visual
representation, I derived a scale from Fischer’s skill theory (1980) and adapted criteria for
identifying various levels of cognitive development through drawings (Table 2, Category 3, on page
62).
Regression analyses, using dummy variable coding, indicated that there were no major statistical
significant differences between the students grouped as English monolingual, Spanish monolingual
or bilingual speakers or as attending monolingual English or bilingual Spanish/English educational
programs. Correlations between scales and expressive processes indicate the reliability of the
measures across the verbal and visual domains in the performances of each group.
It is clear that in examining differences between monolingual and bilingual speakers, researchers
have applied numerous measures of intelligence, creativity, academic achievement, and socio-
metrics. The use of drawing has been relatively rare, if not absent.
My argument is that visual representations hold evidence of children’s levels of cognitive
complexity, particularly when analyzed for their content rather than their graphic skills in
representing perspective. The effects of “bilingualism” and the impact of dual linguistic
development on children’s cognition are imprinted in their drawings. The rational for selecting this
particular age group was that by middle childhood and pre-adolescence, there is evidence of the
manifestation of children’s skills in different domains.
During middle childhood, cognitive acquisitions allow the child to differentiate his/her abilities
across domains (Harter, 1999, p. 9).
In addition, by ages 11-13, most of the bilingual speakers in the study would have been in the
bilingual program for over five years. According to Cummins (1981) it takes over five years for a
monolingual child to become academically fluent in two languages.
By examining the correlation between established frameworks frequently used to map cognitive
development, and the proposed criteria used to identify imprinted cognitive structures as
manifested in children’s drawings, I hope to create a useful scale for analyzing the content of
children’s drawings and contribute a contemporary alternative to traditional research in artistic
development (Eisner, 1967; Eng, 1931; Goodnow, 1978; Golomb, 2002, 1972; Hildreth, 1941;
Piaget, 1983; Lowenfeld, 1939; Luquet, 1927; 1991 Thomas & Silk, 1980).
By examining the emergent patterns that diverse levels of complexity represented in different
mediums and for three distinct emotions, I am eager to reveal information to further our
understanding of children’s expressive processes and confirm the positive results found in two-way
bilingual immersion education in the United States.
As a bilingual educator with a strong background in visual arts, I have consistently considered
drawings to be manifestations of children's cognition. Using art, I have worked diligently to
developed pedagogies that increased children’s mental flexibility and creativity. Children's
contextual cues and structures of representation often reveal their understanding of self, the
nature of their thoughts, and their levels of cognitive development.
In fact, Koppits (1984) said, “During the elementary school years, boys and girls can express their
thoughts and feelings often better in visual images than in words” (p. 2). Selfe (1983), when
reviewing Susanne Langer’s (1953) writings, establishes a direct link between drawing and mental
experiences through the formulation of symbols.
Langer sees drawing as one of the expressions of man’s unique ability to reconstruct reality for
himself through conceptualization or, in her terms symbolization… She demonstrates that a
drawing is a formulation in symbols of the artist’s concepts and mental experiences (p. 23).
Arnheim (1974) also points to children’s ability to communicate in drawings in spite of the technical
challenges of reproducing what is perceived.
Children are technically unable to reproduce what they perceive… Much of the time, however, the
lines are accurate enough to indicate what the drawing is supposed to be like, particularly to the
observer who compares many drawings of the same kind (p. 163).
Literature Review
To further my understanding of methods used to measure children’s cognitive and artistic
development, and to compare monolingual and bilingual speakers’ relative performances in the
various domains, I examined three bodies of literature: 1) models of cognitive development, 2)
artistic development of pre-adolescents, and 3) effects of bilingualism.
Models Of Cognitive Development
Most psychosocial approaches to studying self reconstruct and examine its development through
subjects’ self-reports, which is considered a cognitive construct.
The self is, first and foremost, a cognitive construction... As a result, the self will develop over time;
that is, as cognitive processes undergo normative developmental change, so will self-concepts,
including their very structure and organization (Harter, 1999, p. 8).
In the twentieth century, the progression from theoretical to practical was induced by the
development of tools used to measure and define the personal self. In fact, Sandra Pipp (1990)
describes what has constituted development.
The type of understanding that can be brought to bear on self and other changes with
development, because changes in the structure of the thought influences how people process
information (Case, 1985; Fischer and Pipp, 1984) and so should also influence how people are able
to process information about the self. These cognitive structures serve as filters or upper limits on
processing capabilities (Fischer and Pipp, 1984, p. 253).
James, William (1842-1910)
James (1981) and Jung (1957) presented the theoretical foundation for the study of the self upon
which theorists covered in this literature-review have built. James introduced the idea that the self
is an internal operative entity composed of two parts (Me and I) with various constituents
involving a hierarchical structure (Damon, 1982; Harter, 1998; Fischer & Mascolo, 1998).
James (1983) defined the private, personal (Me), and public (I) selves in pluralistic terms and
proposed three categories for their analysis. These are as follows:
1. Constituents of the (Me) self (which individuals have arranged in a hierarchical scale according to
worth) represent the nuclear part of the Self:
a. Material Self: This self includes our body, possessions, clothing, family, home, etc.
b. Social Self: The multiple social selves are shaped by the roles we play. This is who we are,
how we behave in relation to others and how other see us.
c. Spiritual Self: This is the subjective part of us.
d. The Pure Ego: This involves our soul and the vehicle for the self-measuring process
2. Feelings and Emotions that the four constituents stimulate.
3. Actions encouraged by the four constituents (James, 1983, pp. 280-300).
James (1983) considered the (I) personal self in terms of individual functions and self-knowledge
as a continuously evolving entity involved in the process of awareness incorporating all
experiences, including the spiritual. The (Me) public self was based on societal recognition. He said,
“Man has as many social selves as there are individuals who recognize him and carry an image of
him in their mind” (pp. 282-282). According to Fischer and Mascolo (1998), the multiple selves
have “capacity to represent and regulate the self along multiple pathways, even within the same
individual” (p. 332).
In his Principles of Psychology, James (1983) proposed that “successive perceptions constitute the
mind” (Hunt, 1993, p. 333), and that the self is not a system of fixed relationships but rather an
ever-changing stream of consciousness (Hunt, 1993, p. 159). He also believed in the ability of the
self to “distinguish between the immediate and actual [objective], and the remote and potential
[subjective]” (James, 1983, p. 300).
As his principal method of investigation, James used introspective observation as a means to
revealing the states of consciousness. He explained,
Introspective Observation is what we have to rely on first and foremost and always. The word
introspection need hardly be defined - it means of course, the looking into our own minds and
reporting what we there discover. Everyone agrees that we there discover states of consciousness
(James, 1983, p. 185).
However, revealing the states of consciousness is challenging. According to Damon and Hart
(1982), “it is difficult to observe or characterize a phenomenon that is totally subjective and that,
therefore, may change unpredictably from moment to moment” (p. 845). Their claim was that
James was more of a philosopher than a psychologist, and did not provide a framework for a
developmental approach.
Although James admitted to some individual variation in how the constituents were formulated, he
did not recognize the possibility that their hierarchical interrelations might vary, both across
individuals and within one individual over time (Damon, 1982, p. 844).
According to Harter (1998), the symbolic interactionists Cooley, Mead and Baldwin contributed to
James’ theory by claiming that the self originated in the symbolic imitation of others and that it is
defined by how we internalize what others thought. They proposed that children develop a sense
of self through interaction when they internalize “the actions, thoughts and attitudes of others” (p.
555). Selfe (1983), when discussing Gibson (1979), shares this belief, “The pivot of Gibson’s
theory is that perception results from an interaction between the viewer and his environment”
(Selfe, 1983, p. 25).
Invariably, self- representations portray all aspects of Self (Me and I), involving them in
introspection, which is subjective. For this dissertation, subjects were asked to draw upon
introspections and represent the subjective. Before starting the research process I had to answer
the question of how to measure “cognition” based on the “content” which in drawing has
consistently been thought of as subjective.
My thesis is built upon the assumption that the choice of symbols used to communicate a
subjective message translates it into the objective realm, providing us with concrete agents or
elements, which can be used as data and can be analyzed quantitatively.
Jung, Carl (1875-1961)
Jung introduced the idea of basic character types, suggesting that the self passes through three
basic developmental phases: early, middle, and late life. He suggested that experiences are stored
subconsciously and that art productions reveal the whole self through symbols. Jung recognized
the importance of symbols. He specifically studied visual symbols, both produced and perceived by
his patients, and their application in the psychology of the self.
The self, I thought, was like the monad, which I am, and which is my world. The mandala
represents this monad and corresponds to the microcosmic nature of the psyche… The self is our
life’s goal, for it is the completest expression of that fateful combination we call individuality… [It
is] the individuation process of integrating the content of the unconscious and achieving
awareness of the self (Jung in Jaffe, 1979, p. 78 & 107).
Jung recognized the importance of others in the formation of self, claiming that the self without
relationships could not become whole. He stated,
Individuation has two principal aspects: in the first place it is an internal and subjective process of
integration, and in the second it is an equally indispensable process of objective relationship.
Neither can exist without the other, although sometimes the one and sometimes the other
predominate (Jung in Jaffe, 1979, p. 125).
James and Jung both offered theoretical frameworks that included constituents of the self, which
influenced all subsequent research during this century. Yet, they did not provide us with a
developmental trajectory of how the self evolves.
Nonetheless, Jung encouraged researchers to experiment with unprecedented innovative ideas
and methods, cautioning scholars about the dangers of looking at the self with existing
frameworks, whether scientific, anthropological or psychological. In an introspective moment Jung
declared how he comes up with a paradigm shift,
I must lay aside all scientific knowledge of the average man and discard all theories in order to
adopt a completely new unprejudiced attitude. I can only approach the task of understanding
with a free and open mind, whereas knowledge of man, or insight into human character,
presupposes all sorts of knowledge about mankind in general (Jung, 1958, p. 10).
In relation to drawing, contemporary researchers such as Koppits (1984) affirm that in the content
of children’s drawings, we can find non-verbal cues to the messages they send out about
themselves (p. 2) and that mental representations and events in children’s lives are direct
representations of children’s cognitive development (Gonzalez, 1999, p. 24).
Non verbal representations are perceptually similar to the events that they denote and they are
processed simultaneously. The dual-code theory also assumes that idiosyncratic and context-
dependent
connections between verbal and nonverbal representations exist and lead to a complex associative
network called referential connections (p. 24).
Vygotsky, Lev, (1896-1934)
Vygotsky “argued that culture [and language] actively constructs and largely determines both the
course and the degree of development (Feldman, 1999, p. 3).
The speech structures mastered by the child become the basic structure of his thinking…"there is a
fundamental correspondence between thought and speech in terms of one providing resource to
the other; language becoming essential in forming thought and determining personality features
(http:www.sk.com.br/sk-vygot.html Vygotsky and Language Acquisition, p. 1-3).
Based on the conception of what constitutes the understanding of “a collective” when organizing
puzzles with sets of blocks, Vygotsky postulated 3 phases of verbal thought, each with its steps
(Blunden, 1999, pp. 5-6; Vygotsky, 1994, pp.11-57):
1) Syncretism: objects are organized by their subjective connections
- Trial and error;
- Egotistical Selections made by trial and error based on the visual field; and
- Combinations of both trial and error and egocentric manifestations.
2) Complexes: Child is able to combine the subjective with facts but results are not always
consistent
- Associative Complex – manifested by what is related and familiar;
- Collection Complex – manifested by sets, collections with different attributes;
- Chain Complex – manifested by one object connected to another by common attributes but
that to the next by a different attribute;
- Diffuse Complex – manifested by variability of the attribute which results in an unstable
connection; and
- Pseudo Concept - manifested by the concrete, visible likeness and by the formation of
associative complex limited to a certain kind of perceptual link.
3) Concepts: A single attribute is identified as the basis of a collective
- Potential – in its simplest form
- Concept Proper – concepts emerge when attributes are integrated and abstractions are
formulated
Vygotsky suggests that phases are not clear-cut and that they overlap as children develop. This
idea continues to be scientifically demonstrated in neuropsychological research as development
gets mapped with images of the brain as it develops (Fischer, 1980).
Phases co-exist with each other in the course of development,
predominantly in different areas of activities and at different times and in different situations. At
the same time, each stage builds upon the gains of the earlier (Blunden, 1999, p. 7).
Luria, A. R. (1902-1977)
Luria, a contemporary of Vygotsky, examined the difference between "people with graphic
reflection of reality and those with predominantly abstract, verbal, and logical approach to reality.
He looked at the differences in how individuals abstract and generalize when asked to group and
classify objects. He claimed that perception of self is related to the perception of others.
The perception of oneself results from the clear perception of others and the process of self-
perception reshaped through social activity, which presupposes collaboration with others and an
analysis of the behavioral patterns (http://marxists.org/archive/luria/works/1976/problem.
htm pp. 14-15).
His study was based on the premise that "the essential features of mental processes depend on
the way they reflect reality; therefore a particular form of mental activity should correspond to a
particular level of reflection (http://marxists.org/archive/luria/works/1976/problem.htm, p. 14).
After World War II (WW2) 1939-1945
After World War II (WW2), psychological methodology incorporated to the view that the self was a
cognitive construction defined by external and internal factors. When examining and describing
how aspects of the self develop, contemporary theorists rely on behavioral observations, context,
and emotions, all influencing children’s interactions with others determining their thinking patterns
and cognitive development.
The influence of socio-emotional development changes in children's mental representational
capacity has received increasing attention in recent years (Gunnar & Sroufe, 1991, p. 126).
Socialization experiences in children's interactions with caregivers, peers, teachers, and in the wider
socio-cultural context will influence the particular content and valence of one's self-representations
(Harter, 1999, p. 9).
When describing how aspects of the self develop, contemporary theorists have relied on behavioral
observations emphasizing contextual factors and social influences that influence children’s cognitive
development.
To our understanding of the nature of social experience (or perception), the defining
characteristics of human sociality, distinctive features of the physical and social environment, and
the psychological nature of the family as the primary social structure within which children develop
(Wosniak, 1993, p. 77).
Piaget, Jean (1896-1980)
According to Piaget (1923, 1926), the development of knowledge, thinking skills, and mental
images (or schemes), as well as the construction of logical-mathematical thinking patterns is
comprised of stages.
Piaget’s (1960) stages, which come about by the child’s interaction with his biological and
contextual environments, have served as an enduring framework in developmental psychology and
in other domains.
Piaget argues that these stages in cognitive development are neither biologically programmed nor
given by the environment, but instead constructed or created by children as they actively attempt
to understand the world (Feldman, 1999, p. 59).
The stages are as follows (Flavell, Miller & Miller, 1993, p. 133):
Sensory-motor: (birth - 18 months)
Pre-operational – Symbolic (18 months - 7 years)
Concrete-operational – Relations and Logic of the Symbolic (7-11 years)
Formal Operational – Abstract (11 -15 years)
Piaget claimed that motivation for action or thought is intrinsic to the cognitive maturity or level of
development of the child (Flavell, 1984). After many years of clinical interviewing, noted
observations, demonstration, and experimental evidence of concrete actions that represent the
internalization of mechanical competencies, Piaget (1936) classified four successive stages of
children’s cognitive development. Competencies (Fischer, Bullock, Rotenberg, & Raya, 1993) are
determined by the interaction of skill levels and neural systems
Competence is an emergent characteristic of a person in context… The dynamics of changes in
competence are explained by the analysis of developmental levels of skills as well as a neural
network model (pp. 93-94).
According to this framework, children at ages 7-9 function at the pre-operational and concrete
operational stage. At the pre-operational level, children view the world subjectively. According to
Di Leo (1977), the conceptual way children draw the human figures tells us a great deal about this
intellectual capacity.
Human figure drawings express the child’s subjective view with total disregard of actual body
proportion. Instead there is an emphasis on what is regarded as important – intellectual not visual
realism (p. 36).
At the concrete operational stage, children are able to understand conceptual categories and are
able to understand how others experience the world. According to Di Leo, “concept now
dominates perception” (1977, p. 36). Gardner (1994) says that children at this stage are able to
function at a symbolic level,
[They are] capable of a series of mental operations that constitute a structured whole… The
operations are actions performed mentally; in our terms, the making system is functioning on the
symbolic level (p. 302).
Fischer and Mascolo (1998) established four distinct levels of operation for the self. These are: a)
the biogenetic level which includes the biological functions of the body; b) the organismic agentive
which is the level that includes the person on its own; c) the social-dyadic level which includes all
relationships with one other individual; and d) the cultural-linguistic level which includes the person
as it functions in society at large.
In addition, as children develop various capacities in separate domains (Gardner, 1982, 1983;
Fischer & Mascolo, 1998), some theorists have paid greater attention to the conceptual structures
underlying the identification and meaning attributed to experiences.
The construction of the experience requires the detection and attribution of meaning to
information from the environment. Detection and meaning attributions are dependent of the
development of conceptual structures and processes in which information from and about the
environment is actively gathered, assimilated to appropriate concepts, and thereby interpreted
(Wosniak, 1993, p. 79).
Feldman, David
Feldman (1998) emphasized the distinction between the universal and the non-universal in
development, stressing that both internal and external forces come into play when a child
develops. Feldman (1985) called universal development the inevitable progress that all humans
make as they grow, requiring no special effort, instruction or environment. On the other hand,
non-universal development is the transformation that is highly dependent on environment and/or
individual resources. He claimed that the “movement from stage to stage is not guaranteed in
domains other than universal ones” (1985, p. 84). Feldman (1994, 1999) critiques some theories
of artistic development, Lowenfeld’s in particular. He says that these theories mix universal,
cultural and discipline based factors and suggests that stages of artistic development need to be
reformulated based on current theories of cognitive development.
This dissertation reformulates the traditional measures used in figuring out the “old” stages of
artistic development and identifies “new” criteria based on established Fischer’s (1980) levels of
cognitive development.
Fischer, Kurt
Fischer’s framework has been applied to developmental studies of self and identity development
across cultures (Cheng, 1998; Pipp, Fischer & Jennings, 1987; Karcher, 1995; Kennedy, 1994).
Data on complexity of conceptions of self are collected and these reveal participants’ levels of
cognitive development. In these studies, participants assigned descriptors and numeric values to
their significant relationships, placing them on circular diagrams. Participants grouped descriptors
according to similarities and differences, thereby determining their level of development
corresponding to Fischer’s measure.
In cognitive psychology, Fischer (1980) challenged Piaget’s framework when he discovered that
disruptions in development happen more frequently than Piaget had proposed and that these
disruptions happen in an individual’s performance of a specific skill in relation to context.
Performances take into account the role of context – therefore are domain dependent.
In his dynamic systems theory, Fischer (1980) delineates 13 levels of cognitive development
(Appendix A), overlapping one another at times. He asserts that children develop their cognitive
abilities, thereby internal representations, in incremental patterns progressing from simple to
complex (Harter, 1998, p. 569).
Representational internal working models are maps of the interpersonal environment which
provide individuals with predictions to understand themselves and others... The representational
internal working model is accessed consciously
(Pipp, 1990, p. 259).
Fischer (1980) examines the organization of behaviors in a wide range of skills proposing a
“recurring cycle of four levels: reflexes, actions, representations, and abstractions. He states that
the self undergoes development in different domains, each at its own rate. The self is represented
by its capacities along multiple contexts.
Capacities to represent and regulate the self develop along multiple pathways, even within the
same individual. As such, the self is not a unified entity that emerges at any single point in
development. Rather, the individual consists of a series of weakly connected control systems that
represent and regulate different aspects of self (Fischer & Mascolo, 1998, p. 332).
Skills vary not only between people but also across contexts for a given person (Fischer, Bullock,
Rotenberg, & Raya, 1993, p. 96).
Fischer (1980) describes his levels of development built on predictable changes on the nature of
complexity of thought, which in turn induce action as children learn a skill (Pipp, Fischer & Jennings,
1987; Shaver & Lamborn, 1887; Fischer, 1980, 1989; Fischer & Watson, 1980; Fischer & Pipp,
1984; Fischer & Bidell, 1992, 1997; Fischer & Rose, 1993; 1994; Fischer & Mascolo, 1998).
The cognitive processes through which structures relevant to incoming stimulus information are
accessed, through which that information is assimilated to the cognitive system, are acts of
meaning attribution. The cognitive system is a device for the generation of meaning (Wozniak,
1993, p. 79).
With development, self and other become increasingly differentiated because higher order
representations of categories of action are formed for self and other (Pipp, 1990, p. 255).
According to Fischer (1980) a cognitive structure is “a structure for knowing, a procedure that the
child actively applies to things in order to understand them” (p. 482). He states: “it is in the very
nature of intelligences that each operates according to its own procedures and biological bases” (p.
68). Skill refers to an individual’s capacity to behave and think within a given context and
psychological domain. A skill refers to an action performed at a fixed time in a specific context
(Fischer & Mascolo, 1998, p. 339).
Skills are defined in terms of environment as well as organism… The concept of skill, which itself
connotes a transaction… is a unit of behavior composed of one or more sets… The characteristic
structure of each level is a type of skill (1980, pp. 478-483).
A skill refers to an individual’s capacity to control behavior and thinking within a given context and
psychological domain. A skill is always an action performed in a context (Fischer & Mascolo, 1998,
p. 339).
Competence refers only to the upper limit on the developmental level in observable behavioral
patterns (Fischer, 1980; Fischer & Bidell, 1992, 1997; Fischer & Pipp, 1984; Fischer & Rose, 1993,
1994). Fischer’s approach examines children’s cognitive structures and their development through
self-representations, describing the relationships between two opposite behavioral factors (such as
nice and mean) during pretend play (Fischer, Hand & Hencke, 1993; Fischer & Rose, 1993) or
person and context (Wozniak & Fischer, 1993, p. 9).
Knowledge [is]...a structural mental code embodying information about the environment.
Knowledge is both given in the human biological endowment and extracted from experience
Wosniak, 1993, p. 83).
Children represent themselves in parallel relationships prior to reciprocal relationships, portraying
either positive only or positive and negative tangible features (Fischer & Mascolo, 1998, pp. 360-
361). For example, at the first level of complexity a student describes his drawing with a single
representation as follows: “I am alone in my house. I have my birthday party and I am having
fun. I am dancing and confetti is falling off and balloons are flying all around”.
At the second level of complexity, the representation involved mapping of two several items
focusing on one aspect of each: A child describes himself, his grandmother and his mother in the
kitchen in his house: “my grandma told me that my dog died and my mom is looking at my
brothers to see what they are going to do. There is a path that leads to where they are. I am
thinking how he used to look. He used to live with me in the Dominican Republic with my grandpa.
Grandma is talking with my mom and my mom is looking at my brothers are thinking that she may
have to say, excuse me I have to go check on my brothers. My grandma is eating next to the
microwave and there is a fridge, the sink stuff and a window that you can see outside”.
At more complex levels of interaction, children have underlying intentions motivating their actions.
This involves more complex systems of operation as children develop behavioral expectations of
others (Fischer, 1990, pp. 5-6).
Interactions differ from other actions in the nature of the expectational control that the cognitive
system exerts as the interaction unfolds (Wosniak, 1993, p. 82).
At the third level of representation the child involves relating several aspects of each item in a
system. He describes herself and grandmother in Cambridge at the funeral home: “I am crying
because my grandma died. I am thinking it was my fault because we went places when she did not
have a coat on and it was cold. She was really sick and we still went out”. Finally, at the fourth
level of representation, the child coordinates several concrete inter-relationships (systems) in terms
of an abstract or intangible characteristic. For example, a child draws himself alone at his house
when he was locked out: “I am waiting for my brother to come because I forgot my keys and
sometimes it takes a really long time so it makes me really mad because he is supposed to come
home early. He takes my keys everyday when he is not supposed to because he lost his own keys.
He gets up earlier than me and I think I can’t find the key, but it is that he has it. I am thinking,
when is my brother going to be home?”
Exploring these approaches further, in this dissertation Fischer's framework has been examined as
it applies to drawings by allowing subjects to draw their relationships.
Application Of Fischer’s Framework To The Proposed Study
Experiences get translated into meaning and in order to represent those meanings to communicate
with others, humans create symbols. Symbols are individual in their nature, but collective in their
use.
In addition to experiencing the world, human beings also generate verbal and imaginal symbols to
stand for it. Semiotic activity, the mental activity of symbol generation, has both representational
and cultural functions. Representationally, symbols stand for something other than themselves.
Through symbol, we can bring to awareness objects and events not currently present to the
senses.
Culturally, symbols embody the historically derived system of social meanings held in common by
members of the broader society. Much of human ability to communicate interpersonally, depends
on the fact that we possess shared systems of symbols (Wosniak, 1993, p. 80).
Fischer's framework (1993) is well suited for my study because of the symbolic congruence that
exists between play and drawing, and its potential to contribute to our understanding of
development levels in children’s drawings. In both activities, children rely on the interaction of
perceptual and conceptual processes creating a symbolic representation that parallels reality
(Krampen, 1991). Piaget (1983) also correlated play, which is based on imitation of real life
situations, with the formation of symbols.
Symbolic play… always involves an element of imitation functioning as the significant, and early
intelligence utilizes the image in like manner, as a symbol or significant (p. 126).
Fischer’s established framework to map skills development can be used to measure the nature of
represented interactions and relationships in drawings, especially since developmental sequences in
drawing that have been proposed for pre-adolescence (Di Leo 1973; Golomb, 1972) are not
specific enough to describe the types of disruptions identified by Fischer and colleagues in cognitive
psychology and neurology. Based on Fischer’s developmental progression, children are gradually
able to distinguish social roles and their complexities. Children begin with the understanding of
relationships that are similar in nature, followed by understanding interactions that are opposite,
sequential, reciprocal, instrumental, and collective.
Skill theory provides a suitable developmental system for studying children’s cognitive profiles
because Fischer (1980) claims that development is quantifiable even if multifaceted and divergent
according to domain and individual differences. This study explores how the representation of
cognitive structures can be quantified and interpreted into levels of complexity deduced from
children’s visual and verbal symbols.
WATSON, Malcolm.
Watson (1990), in his analysis of role-play proposes parallels steps to Fischer’s levels. He describes
the rules of dynamic transformation of interaction in particular when children enter the age of
representations, when they are able to formulate their actions according to social motivation and
to coordinate various roles.
A social role is defined as the cluster of prescribed and expected behaviors that pertain to a
particular category of people. The cluster of behaviors for one role is determined by the
relationship of that role to complementary roles. Roles are complementary when they are normally
associated and functions of one cannot be described without reference to the other (Watson,
1990, p. 271).
Watson’s steps begin with the differentiation of agents rooted in concrete behaviors (comparing
similarities and differences between agents) and progress to the integration of the roles based on
social relations and eventually on networks of relations in preadolescence. Like Fischer, Watson
points out that "it is important that the tasks be related to the specific domain in which one is
testing skill development" (p. 274).
Watson’s model is useful in providing details that I considered in the analysis of drawings seeing
that his steps 2-6 correlate with Fischer's levels of representational systems:
Step 2: Self as Agent - Social Role (Corresponding to Fischer’s Rp1) – Foci of behavior is on
the relationship between two complementary roles;
Step 3: Passive-Other Agent - Shifting Social Role (Corresponding to Fischer’s Rp2) – Child
is able to understand that one person’s role can shift, but does not understand that someone
could occupy the two roles simultaneously;
Step 4: Active Other Agent - Role Intersection (Corresponding to Fischer’s Rp3) – Child is
able to integrate the roles of self and others by bearing in mind multiple and simultaneous role
relations; and
Step 5: Integration of Roles - Role Network (Corresponding to Fischer’s Ab1) – Child is able
to compound several role intersections and can think about combining them to form a network;
and
Step 6: Network Relations (Corresponding to Fischer’s Ab2) – Children are able to relate
one network with another.
To sum up, children first imitate the actions of others, portraying emotions as either positive only
or negative only prior to combinations of positive and negative (Fischer & Mascolo, 1998, pp. 360-
361). At the next level, children are able to oppose, remedy, and reciprocate the action of others,
as well as use others as instruments to perform self’s desired actions. As they are able to count on
others to perform what the self cannot accomplish independently, they get involved in role
intersections where self is part of a compound action and others are recipients. Based on this
system of role intersections, and network relations, self becomes aware of shifting social roles
(genetic, physical, and emotional), where self is seen as an agent of actions that others may or may
not complement or support. Gradually, children begin to grasp abstractions as they develop
deeper understanding of collective actions. Finally, through abstractions, they are able to express
social or personal motives through acceptable symbols, which generally occur during adolescence
(Cicchetti & Beeghly, pp. 267-279). At these more complex levels, children’s actions manifest
hypothetical ideas such as justice, conformity, intimacy, etc. (Fischer, 1990, pp. 5-6).
Fischer and Mascolo (1998) summarize the overall agreement on the theories of self when they say
that the self develops as the physical body interacts with family values, society, culture and
historical context.
Selves are not decontextualized, abstract, or disembodied entities. Rather, selves are constructed
in particular bodies within particular physical, social, cultural and historical contexts. The
production of affective and bodily experiences consists of an especially important source of the self’
s embodiment (p. 350).
The self develops influenced by both internal and external factors. The development of self can be
observed, measured and analyzed. As methodology evolves research has the capacity to be more
accurate and reliable. As long as children’s construction of their self-concept is measurable
(starting simple and moving to complex) and that the representations of self reflect levels of
complexity in the emerging patterns, it is feasible to compare two groups of children with distinct
cognitive experiences.
The design of the drawing tasks used as data for this dissertation included representations of the
self in its physical and emotional contexts associated with the lives of children. My hope is to
contribute to further our understanding of cognitive development through children’s visual
representations of self and others.
Artistic Development Of Pre-Adolescents
Nature Of Drawing
Across the board researchers state that children’s drawings represent the interpretations of
reality. Many believe that data drawn from drawings provide a source of understanding of children’
s psychological and emotional development.
Young children are not only indulging in self-expression in their drawings, but are representing real
relationships. Their drawing
can be checked against structural descriptions of the external world (Freeman, 1980, p. 37).
Drawings capture symbolically on paper some of the subjects’ thoughts and feelings. It makes a
portion of the inner-self visible… More is revealed by the content, which is largely determined by
the way the subject, consciously or unconsciously, perceives himself, and significant other people in
his life (Klepsch & Logie, 1982, p.6).
Drawing is, above all, the construction of meaningful patterns out of simple abstract elements
(Freeman, 1980, p. 55).
Freeman (1980) explores what constitutes a representation and identifies three components: a
definable object, definition of a relationship, and the study of action. “Drawings clearly represent
objects and people, and the relationship between them, and they got through a well defined
process of development” (p. 50-51).
The criterion for selecting the appropriate inclusive task in drawing was that drawing must capture
both process and product as well as reflect internal and external structures. This approach is well
supported by some researchers.
The very act of drawing involves many interacting processes, sensorimotor coordination,
perception of the external model when there is one, consulting inner image or schema or
representation in memory when no external model is present, production of a graphic form and its
correction, understanding of pictures a medium of representation and communication (Lange-
Kuttner & Thomas, 1995, p. 2).
Arnheim (1974) postulates that drawings are inclusive of both internal and external structures.
In thinking about why art work proceeds from the simpler to the more complex level, we realize
that internal as well as external factors must be considered. Internally, the organism matures, and
as it becomes capable of more differentiated functioning, it develops the urge to apply this
capacity. The development, however, is not conceivable without the external world, which offers
the whole variety of directional relations (p. 188).
Yet Arnheim, in Harris (1963), recognize the value of drawings as final products described as
voluntary statements about precise realities and particular characteristics.
The work of art is a statement about the nature of reality… From an infinite number of possible
configurations of forces, it picks and presents one… the final solution of the problem of how to
organize a reality pattern of given characteristics (p. 177).
Other researchers, such as Basset (1977) suggest that drawings are not the reflections of internal
representation. Basset designed a study to demonstrate that children’s concepts of self are
complete in spite of what their representations tell us because young children (4.9 years old) were
asked to draw a human figure, and then to construct a figure with given parts. Children most
often drew incomplete pictures, yet never missed using all the body parts to build a complete
figure. Basset explained that “the image is complete, but individual components are not clearly
differentiated” (p. 49), which led her to conclude that drawings do not represent children’s internal
representations. When scanning a picture, young children do it along a vertical axis, paying
attention first to top and then to bottom (pp. 58-59).
Drawing As Representation
Drawing, when intended, can reflect the artist’s reality of what is perceived and conceptually
understood. As verbal language develops from simple to complex, visual arts develop from
concrete to symbolic. When external stimuli interact with the internal mechanisms, they produce
concepts, relationships and systems. These get expressed through narratives, dialogue and
symbolic representations. Drawings are different from photographic images in that they reflect
concepts based on the perception of reality
According to Arnheim, children tend to develop concepts before they are able to find means of
representation.
Image making of any kind requires the use of representational concepts. Representational
concepts furnish the equivalent, in a particular medium, of the visual concepts one wishes to depict,
and they find their external manifestations in the work of the pencil (Arnheim, 1974, p. 169).
Freeman (1980 separates the process in two: projection and denotation systems. These are based
on spatial axes (linear perspectives or points of view), as he asserts that, “drawing gives a printout
of his conceptual store… that the child’s drawing is knowledge dominated… [and] that the child
can draw far more sensitively to visual models than appears, because he extracts the ‘structural
core’ of the model” (Freeman, 1980, p. 28).
Freeman (1977) suggests that observing process and sequence in which drawings are made is
more valuable than the final product because the final product tells us little about the child.
Convincing though post hoc analyses of finished products may be, they cannot substitute for
performance analysis of ongoing processes (p. 25).
Arnheim (1974) also emphasizes the importance of process as a component of research that
should not be neglected,
The order in which children produce the various parts of their drawing is most relevant to the
psychological meaning of the work and should not be neglected in research (p. 173).
I strongly disagree with Freeman’s statement that the finished product provides little information
about the child, yet I acknowledge the value of process. On the other hand Freeman’s
recommendation for designing research studies using art as a data collection tool is well
appreciated.
An image can be made to work if one devised a task, which gives the image a functional role to
play… One can devise studies in which people behave as though they organized their behavior
around some analog representation of a think in the outside world (Freeman, 1980, p. 42).
When looking at drawings as finished products it has been prevalent to look at the whole picture,
Karen McNiff (1982) declares that children communicate visually and that art represents a
synthesis of a child’s world beyond imitation.
It is not a process of imitating or copying the physical world, but rather of synthesizing life
experiences. Art then becomes a means through which the child can communicate about the
phenomena which are too complex to describe verbally, but which are being perceived and
integrated into the child’s organization of reality (p. 275).
The system used to analyze drawings in this dissertation starts with the examination of the finished
products in terms of the evident relationships of the parts that constitute that “synthesis” and
evaluate the nature of the relationships between those parts. Maturity is represented by the
degree of integrity with multiple layers (or systems) of relationships. When looking at students’
graphic compositions, the relationships are based on elements of perspective (point of view).
When examining visual content and verbal narratives, relationships are based on the nature of the
depicted interactions of self.
Drawings Of Self And Significant Others
The analysis of drawings of self and significant others was first introduced by Hulse (1951) [as
reviewed in Klepsch & Logie (1982, pp. 77-110)] with the “Draw-a Family” technique. Norman
Freeman (1980) defines self-image in drawings as a concept that integrates information from
diverse sources and defines choices about spatial location and actions.
The body image is supposed to be a sort of composite conception of one’s ‘self’ constructed from
information from different sources… Many people are convinced of its usefulness though,
since it is an appealing notion that drawings are connected with one’s conception of, and feelings
about, oneself. This could then provide a link between the representation of the self in highly
structured drawing and expression of the self… The body image [is] a construct used to explain
how we are enabled to take decisions about spatial location and the actions, which are necessary in
dealing with that information (p. 45).
Attention is paid to the degree of synthesis examining: closeness, positions, omissions of figures,
size relationships, etc., to determine degree of maturity, family dynamics, as well as cultural and
racial differences. Using the same approach, Hildreth (1941) studied a sample of 149 train
drawings, produced by children ages 2-11, for their graphic narration mostly achieved through
symbols and their story telling arrangements (p. 159). She found that “bright children tend to
show in their drawings more correct synthesis than others, and older ones more than younger
ones” (p. 151). However, she claims that development is not simply about synthesis, but rather it
is about “ progressing more and more toward creative, imaginative drawing, using habitual
schemas but constantly introducing original touches (p. 153). With the recognition that visual
representations need not to replicate all aspects of an object, omissions lose their influence as
measure of development. This type of analysis is frequently used today as diagnostic aids in
psychological testing (Kramer, 1971, McNiff, 1992) rather than to map cognitive development.
Theories that examine drawings as finished products originate with the idea that drawings are
projections and intimate representations of the self (Harris, 1963; Machover, 1949).
The figure drawn is related to the individual who is drawing with the same intimacy characterizing
that individual’s gait, his handwriting, or any other of his expressive movements (p. 43).
Klepsch and Logie (1982) wrote about the four projective uses of drawings; one of which reveals
how children perceive themselves in relation to others. They describe these drawings in the
following way,
Group drawings are useful if one wants to find out how a child perceives himself within a particular
group. When children draw themselves along with their family or friends or teacher or
schoolmates, they project into their drawings their view of themselves in relation to others in the
group (p. 12).
Stages Of Artistic Development
There is no controversy over the facts that the visual arts comprise of systems of meaning with its
own rules and developmental sequences; that drawing production is directly linked to mental
development (Arnheim, 1954; 1974; Di Leo, 1977; Eng, 1931, 1957; Golomb, 1972, 2000;
Goodenough, 1926; Harris, 1963; Kellogg, 1969, 1970; Klepesch & Logie, 1982; Koppitz, 1968;
Lowenfeld, 1939, 1957, 1970).
Artistic development was originally built on observation of changes that individual children
demonstrated as they invented ways to express their ideas in drawing.
Drawing development is largely accounted for by creating children’s discovery or invention of the
pictorial schemata and of procedures for creating these schemata on paper (Thomas, 1995, p.
107).
The documentation and interpretation of events and behaviors during children’s acquisition of
graphic, symbolic, and communicative competence has facilitated the demarcation of predictable
stages of artistic development (Eng, 1931, 1957; Davis, 1997, 1991; Di Leo, 1973; Duncum, 2000;
Fischer, Jennings, Lamborn, Pipp & Shaver, 1985; Gardner, 1980; Golomb, 1972; Kellogg, 1966;
Koppits, 1984; Korzenik, 1972; Lowenfeld, 1939; Matthews, 1999; Thomas & Silk, 1990; Varon,
2000; Winner, 1982,1986).
There is also widespread agreement that in the image lies in a conceptual understanding. This
understanding (therefore image) is simply expressed at the beginning levels and gets more
complex with experience and skill development. Consequently, a collection of drawings from
children at various ages and skill levels provides a developmental continuum, and each drawing a
snapshot of children’s thinking, preferences and skill level.
According to Harris (1963), Hevner (1935) and Walton (1936) were the first to demonstrate how
drawings capture and communicate children’s emotions developmentally and to compare this
communication to the structure of spoken language. “Lines and forms may express feelings as
stated in adjectives” (p. 39). Harris (1963) explains Walton’s finding that “children’s attribution of
feeling involves principally including the pleasant-unpleasant dimension. However, as they learn
more words and concepts, children make finer distinctions and discriminate more polar dimensions”
(p. 39).
Representation implies abstraction and simplification (Arnheim, 1954) as opposed to duplication.
Duplication demands precision, attention to detail and to multiple relationships, a developed special
framework, and the concept of measurement (Piaget & Inhelder, 1956). Progress in the ability to
copy geometric figures parallels to some extent the increasingly stringent demands of “likeness”
which the child sets for himself, advancing from the early simple forms of equivalence to more
complex and mutually regulated ones (Golomb, 1974, pp. 184-185).
Golomb (1972), agreeing with Arnheim(1974), affirms that children’s representations in drawing
evolve from simple to complex that visual language operates similarly to verbal language. She says,
Drawing is a uniquely human activity whose complex syntactic and semantic development can be
studied systematically… Arnheim calls attention to the general developmental course that
proceeds from simple to complex forms, a developmental principal of differentiation that
characterizes all cognitive processes (pp. 2 & 31).
Golomb (2002) explores the nature of representation and the role of internal models of mental
images. She establishes the difference between development that moves towards realism and the
one that moves towards the representation of diverse objects (p. 5).
Golomb (2002) also differentiates between representation and perception, expanding Piaget’s view
that representations are systems of differentiated symbols that stand for recognizable objects.
Her broader view explains representations as “inventions of pictorial forms that stand for the
intended objects” (p. 5). It is the mechanics, principles, and symbols children use to express their
ideas and emotions about themselves and others that are considered projections of what is in a
child’s mind. Development progresses from “topological relationships that ignore true shape and
proportion” to realistic depictions, intellectual realism and finally, photographic realism (p.11). Di
Leo (1983) implied the same view of development when he says, “These terms express, in
substance, a metamorphosis in thinking from egocentricity to an increasingly objective view of the
world” (p. 37).
Freeman and Cox (1985) propose a completely different view of development. They claim that
development moves from object centered to view centered.
Goodenough (1926) was the first scholar to demonstrate that through the study of drawings of
the human figure (HFD) one could measure children’s intelligence and examine the salient
characteristics of a child’s personality. Her measure (which was later revised to be the
Goodenough- Harris measure) provided many scholars with objective criteria with which to analyze
drawings, propelling numerous comparative studies of children’s drawings (Harris, 1963, p. 11).
As for the development of mental images, Piaget (1923) and Luquet (1979) identified three and
five stages of development respectively. Krampen (1991) combined Piaget and Luquet’s phases
and created the following four developmental sequence:
1) Synthetic Incapacity or Scribbling (ages 1-3)
2) Fortuitous and Failed Realism (ages 3-5)
3) Intellectual Realism (ages 5-8)
4) Visual Realism (ages 8-12) (p. 40)
Consequently, children ages 7-9 who have a tendency to draw at the Intellectual Realism Stage,
draw individuals according to their internal mental models, which match the subject whether the
person is posing or not. Krampen (1991) reveals details of what is in a child’s mind and how it
translates into the visual language when drawing in the Intellectual Realistic Stage,
Intellectual realism refers to the fact that children draw everything they know even if it is not
actually visible from a certain vantage point…. Proximities are correctly rendered: thus arms are
attached to the trunk and eyes are placed side by side in the head (even if this is shown in profile).
Order is present in the drawings of landscapes or houses, although no yet in accordance to
coordinates (e.g., the chimney is attached to the house perpendicularly to the roof). Surrounding
and enclosure become so important that in many instances the inside of an object (e.g., a house) is
shown together with its surrounding outline in so-called transparencies. Some projective and
metrical relationships of space are shown, but in an incoherent fashion. Since there is no
coordination of viewpoints or perspectives, a single drawing in this phase may present many
different viewpoints at once (pp. 39-40).
Lange Kuttner, Christine and Reith, Emiel.
Lange-Kuttner and Reith, (1995) base their levels of development on the changing
“conceptualization of objects and space” by examining three components: objects and emptiness,
graphic configurations, and measurement. They claim that as a child develops he acquires
reversibility of thought allowing him “to deconstruct and construct, to structure and measure both
real objects and graphic forms” (p. 87). The subject becomes able to modify “simple geometric
forms into irregular complex shapes, or the supplementation of human figures with spatial axes
systems” (p. 91). Their developmental sequence progresses from global to articulated drawings.
Lowenfeld, Viktor.
Lowenfeld (1939) revisited Piaget’s theory of development in children’s drawings and proposed a
framework for analysis organized thematically. Lowenfeld affirmed that art was an indicator of
children’s thinking as well as the expression of their emotions. He demonstrated that successive
drawings on a theme reflected development.
“Lowenfeld believed that children’s growth through art was analogous to the process of organizing
thoughts and the development of cognitive abilities (Malchiodi, 1998, p. 66).
He looked at creativity in subject matter from the point of view of expression of form, space, and
color, based on the premise that “all artistic ability has its roots in universal human dispositions. It
is determined only by the exceptional intensification of the abilities of a certain individual sense” (p.
5).
Lowenfeld (1939) proposed six stages of development, four of which I use in my study. He
delineated his stages based on observation of children’s use of artistic elements such as color, line,
etc. and their role in the expression of emotional meaning. His six stages of development are:
1. Scribbling Stage (2-4 years)
2. Pre-schematic Stage - First Attempts at representation (4-7 years)
3. Schematic Stage - (7-9 years) Definite form concept.
4. Drawing Realism (9-12 years) Symbolic use of form
5. Pseudo-naturalistic Stage (12 years) Reasoning
6. Period of Decision (p. 27).
Under Lowenfeld’ schematic stage children are able to represent concepts and emotions in their
drawings, performing according to their developmental capacity and their ability to manifest their
insights about their self and others in drawing.
Lowenfeld (1952) emphasized the individual’s ability to express his or her unique characteristics
rather than objective reality (Harris, 1963, p. 38). Lowenfeld said “a child will emphasize and
exaggerate on his drawings those parts of the figure, which have special meaning for him”
(Koppits, 1968, p. 94).
Lowenfeld’s (1939) works in the 30’s and 40’s stressed the need to find links between art and the
psychology of children’s personality development. He said, “What is lacking above all is a
psychological analysis of the correspondence between the development of pictorial art and the
total development of the child” (p. 61).
Lowenfeld’s (1939) framework for analysis of drawings is organized thematically and it identifies
graphic indicators (Di Leo 1973; Golomb, 1972; Lowenfeld, 1939). His developmental sequence is
based on his observation of children’s use of artistic elements (such as color, line, and perspective),
which approximately corresponds to Piaget’s stages of logico-mathematical thinking. It is the lack
of empirical data that propelled Eisner (1967) to criticize Lowenfeld’s scale.
It contains numerous assertions that lack adequate documentation. The stages that are described
are not the result of empirical studies using scientific controls to insure objectivity but insightful,
even if dogmatic conclusions drawn form years of experience working with children (p. 14).
In spite of this criticism, Lowenfeld (1970) successive stages of development for drawing have
been the most widely used for the analysis of elementary school children’s art. This is why I choose
to use it in my study as one of the rating systems to identify levels of complexity in graphic
composition: Pre-Schematic (ages 4-7); Schematic (8-9); Realism (9-12 years), and High Realism
(12-14) (pp. 36-40). These stages may correlate with the developmental disruptions identified by
Fischer (1980).
Although Lowenfeld’s (1939) stages include ways to identify the art of adolescents, the predictable
nature of drawings produced by children older than 12 years of age is less certain (Davis, 1991). In
fact, adolescents may not reach the last stage of artistic development “unless they continue or are
encouraged to make art (Malchiodi, 1998, p. 67).
The adaptation of Fischer’s scale of cognitive development proposed in this study offers a
compelling alternative to Lowenfeld’s scale which is out-dated and does not take into account the
content and meaning of children’s drawings.
Developmental Sequences Of Graphic Representation
In the largest attempt to put together a comprehensive sequence of graphic representation,
Rhoda Kellogg (1966, 1969, 1970) greatly influenced by Goodenough, created a comprehensive
pictorial continuum of the development of children’s drawings of the human figure from a
quantifiable perspective. Her stages of development based purely on the graphic characteristics in
children’s representations were derived from thousands of scribbles and drawings of the human
figures she collected from around the world. Kellogg, however, limited her collection to drawings
from children in the pre-school years.
Goodnow, Jacqueline.
Goodnow (1978) provides descriptions of devices children use in order to show human figures in
movement. She proposes a developmental sequence that includes 15 features, which include
increasing details, dimension, overlaps, manipulation of body axis and bending or flexibility.
Golomb, Clair.
Golomb (1972), author of The Child’s Creation of a Pictorial World, began to look at elementary
school-aged children’s development of graphic representation pointing to their conceptual nature.
Her work focused on “the principles that underlie the graphic mode of visual symbolization” (p. 3).
She observed the common themes and gender preferences that children exhibit at certain ages
and explored the individual and cultural variations on those themes (p. 6).
Golomb (1972) used age and grade levels as ‘convenient ordering devices’ and defined stages of
development as “changes that are largely propelled from within. Stages are determined by the
child’s own and active experimentation with the medium, as well as by the nature of the objects
she wishes to depict” (p. 6).
Golomb contributed to the perspective that competency development is domain specific and that
art is a cognitive activity. In his theory of Multiple Intelligences, Gardner (1983) suggests eight
domains in which the self can develop competencies. With reference to art, Gardner says that it is
an activity of the mind, rather than that of the senses. Students perceive, think, feel and then
express as they produce art. Gardner looked at children’s drawings from a developmental
perspective also organizing them in thematic frames.
Gardner, Howard.
In his book Artful Scribbles, Gardner (1980) argues that children’s drawings may exhibit some of
the same qualities of established artists, and determined that artistic development begins with
scribbles and ends with a search for realism. Gardner proposed a framework for aesthetic
development based on Piaget’ stages. Gardner’s (1980) work focused mostly on development of
aesthetic perception, asking children to report on their ideas about works of art displayed to them.
Willats, John.
Willats (1995, 1997) conducted a study of ways that children represent a table with different
objects on its top. According to his findings, the changing shapes chosen for the representation of
their specific point of view and the foreshortening mark development.
Davis, Jessica.
Davis (Davis, 1991, 1993, 1997; Lawrence-Lightfoot & Davis, 1997) who looked at representational
and non-representational artworks done by children and adults was interested in vehicles and
strategies individuals use to express three specific emotions: happiness, anger and sadness. She
demonstrated the theory of U-shaped development (Davis, 1991, 1997) in terms of expressive and
affective characteristics of the three different emotions. In addition to the expressive qualities of
the collected art work, Davis analyzed the content of the formal artistic properties such overall
expression, overall balance, appropriate use of line and appropriate use of composition. She found
that children’s facility with graphic symbolization most often submerges by age twelve. Children
ages 7-9 work in what is called the “literal stage” and are appropriate candidates for the
production of detailed interactive drawings.
Emphasis for analyzing very young children’s drawings include the study of scribbles (Kellogg,
1966, 1969, 1970; Gardner, 1973, 1994, 1980); as well as representations (Eisner, 1967; Di Leo,
1973, 1977, 1983; Eisner, 1988; Davis, 1991, 1993, 1997; Lawrence-Lightfoot &Davis, 1997;
Goodenough, 1926; Golomb, 1972; Goodnow, 1977; Goodman, 1978; Krampen, 1991; Varon,
1997); introspections (Jaffe, 1979; Lowenfeld, 1939); and retrospectives (Eng, 1931, 1957;
Fischer & Watson, 1980; Varon, 1996, 2000).
Comparative Studies
Many comparative studies emerged and numerous techniques were developed to study children
with specific problems using Goodenough & Harris basic way to categorize and analyze children’s
figure drawings. This technique has been mostly used as diagnostic aids in psychological testing (Di
Leo, 1973, 1977, 1983; Klepesch & Logie, 1982; Koppitz, 1968, 1984; Kramer, 1971, 1993; McNiff,
1992; Peterson & Hardin, 1997).
Comparative studies using control groups include those looking at race, gender, socio-economic or
cultural backgrounds, behavioral, physical, or mental abilities groups, etc. These studies noted the
differences in graphic treatment and occurrences of such treatments by the particular populations
(Eisner, 1967; Klepesch & Logie, 1982; Koppitz, 1968, 1984; Lange-Kuttner & Edelstein, 1995).
In her cross-cultural study, Koppits (1984) found that Argentinean boys and girls drew more
figures in action than American children (p. 72). Koppits did not report explicit details about
children’s linguistic experiences and attributed the difference to children’s cultural experience.
Koppitz, Elizabeth.
Koppits’ (1984) cross-cultural comparative study finds that Argentinean boys and girls draw more
figures in action than American children (p. 72). Koppits attributes this difference to children’s
cultural experience.
Eisner, Elliot.
Eisner (1967) makes reference to the relationship between drawing stages and language
development, and states “…the linguistic system that an individual learns serves to structure his
perception and conception of the world” (Eisner, 1967, p. 34). Eisner’s (1967) report, A
Comparison of the Developmental Drawings Characteristics of Culturally Advantaged and Culturally
Disadvantaged Children, documents that culturally disadvantaged children were lagging behind in
development, “it takes four years for the disadvantaged group to perform at a level that
approaches the culturally advantaged first graders” (p. 72). When he compared developmental
characteristics in the drawings of culturally advantaged and disadvantaged children (from grades
one, three, five and seven), he found that culturally disadvantaged children were lagging behind in
development.
Eisner set up a scale that correlates with cognitive development to assess stages of development
found in children’s art. He compared two groups of students and examined the relationship
between children’s language and artistic development. His scales constitutes of 14 categories
“directly related to the presence or absence of a horizon line or lines and the placement of
morphemes” (p. 35). Consequently, Eisner’s scale is of the same nature as Piaget and Lowenfeld in
that it measures graphic composition or “aesthetic syntax”. If graphic composition measures
aesthetic syntax, I could claim that my proposed visual scale measures “visual semantics”.
Goodenough, Florence.
Goodenough (1926) found that girls represented differentiated dimensions of the human figure
with more details than boys; and boys depicted movement more frequently than girls.
Lange-Kuttner, Christina & Edelstein, Wolfgang.
Lange-Kuttner & Edelstein (1995) conducted a longitudinal study of social factors to graphic
competence and human figure concept. They claim that the transformations of children’s cognitive
development are shaped by their social realities. They found that only drawings of the human
figure were impacted by social variables. “The differences in depiction of volume and movement
varied both as an effect of social class and gender” (p. 163).
Gender differences in drawing do not merely relate to the biological differences between the
sexes… The present study shows that girls were more likely to differentiate the human figures
while boys were more often reduced complexity or depicted movement (p. 169).
Krampen, Martin.
In his comparative study Krampen (1991) looked at cultural differences in the drawings of two
different kinds o buildings. He studied Turkish and German children from urban and rural areas.
He found no difference between the urban children and between the rural ones. Urban children
distinguished more clearly between types of apartment buildings and residential houses than village
children.
In short, research on the development of figure drawing has examined children’s mental and
emotional capacities and development across age, gender, race and cultural experiences. Such
studies have provided in-depth understanding of personality and behavioral aspects of the
development of self in relation to others; and they confirm that children’s drawings are in fact a
reliable measure of development.
The implication of these findings for my study is the confidence that drawings will provide reliable
evidence of children’s cognitive structures, representing their developmental levels, which in turn,
may also be compared. It is also clear that the question still remains as to how cognitive
competency and development in one domain has impact upon another.
Comparison of drawings produced by bilingual and monolingual children, two groups with radically
different cognitive experiences, is unusual. Few studies have investigated the possible effects of
bilingualism on children’s artistic development. Given that linguistic experiences have significant
cognitive impact, the proposed study may provide insight into the effect of bilingualism in children’s
cognitive development.
My study is an attempt to cross the historic boundaries that have existed between different
domains, and is the first one comparing the drawings of monolingual and bilingual children for their
representative characteristics, exploring interaction between graphic composition, language
development, and the visual content.
My intent is to examine drawings of children ages 10-12 because they seek “to make substitute
images of selected aspects of the world about them. However abstract their representations,
children’s purpose [in drawing] is to create graphic equivalent of something perceptually
observable” (Duncum, 2000, p. 44). Often they include scenes of themselves interacting with
others (Davis, 1991; Lowenfeld, 1939; Varon, 2000). I researched frameworks currently
used to identify levels of artistic and cognitive levels of development and explored how these two
perspectives can inform each other.
Bilingualism
There are far more studies that cite the advantages of bilingualism than those that point to its
negative effects.
There are close to 150 empirical studies carried out during the past 30 or so years that have
reported a positive association between additive bilingualism and students’ linguistic, cognitive, or
academic growth (Cummins, 2000, p. 37).
The debate on the effects of bilingualism dates as far back as 1935 when Vygotsky (1962)
suggested that bilingualism alters an individual’s abstract thought processes. Cummins (2001)
discussing the opposite view says, “The Piagetian School has consistently held that the
development of the basic cognitive schemata owes little to language.” (p. 20). Piaget (1967) in
fact, stated that cognitive growth is hardly affected by bilingualism.
The continuous supported acquisition of two languages promotes cognitive, linguistic and
academic growth (Collier, 1987; Cummins, 1981; Hakuta, 1986).
As word meanings and concepts develop. They become increasingly embedded in a system or
network of logical and semantic inter-connections (Cummins, 2000, p. 61).
The process moves swiftly from the beginning level to oral fluency found in natural language.
Academic fluency in two languages (Academic Language Proficiency CALP or (bi-literacy) however,
takes five to seven years. CALP is essential in order to attain full cognitive benefits of bilingualism
(Cummins, 1981).
It reflects the registers of language that children acquire in school and which they need to use
effectively if they are to progress successfully through the grades. The distinction [between natural
and academic language] highlighted the range of cognitive demands and contextual support
involved in particular language tasks or activities (Cummins, 2000, p. 59).
In a recent publication, Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001) found that
bilingual and monolingual children develop at similar rates in biologically controlled mechanisms
related to linguistic acquisition. Bilingual children achieved the same milestones (for example,
number and types of utterances) developmentally the same as monolingual children. Petitto &
Kovelman (2003). found that bilingual children performed better than monolingual children in tasks
that involved abstract reasoning and attention. Petitto (2003) found that bilingual children
performed better than monolingual children in phonological tasks. In addition, through the
analysis of sequential or simultaneous language mixing tasks, they found that bilingualism did not
create confusion.
Negative Effects of Bilingualism
Few studies maintain that bilingualism causes cognitive deficits such as: mental confusion (Saer,
1923), linguistic handicap (Pintner & Keller, 1922), phonic interference (Weinreich, 1957), semi-
lingualism (Skutnabb-Kangas & Toukomaa, 1976) and lower score on verbal ability tests, one of
the most common findings (Ben-Zeev, 1977, Smith, 1933). These studies have been criticized for
their methodology: one such criticism is the lack of matching control groups (Lambert, 1972); and
another is that bilingual participants in these studies had been subjected to subtractive educational
models of bilingualism (Cummins 1976).
Thus, the early findings of negative cognitive effects associated with bilingualism would be
explained by the fact that minority language children in these studies often failed to develop
sufficient high level of proficiency in the school language to benefit fully from the educational
experience (Homel & Palij, 1987, p. 60).
Positive Effects of Bilingualism
Studies putting forward that bilingualism produces general advantages in cognitive abilities are
quite numerous. Findings commonly cite that bilingual children show greater right hemisphere
involvement in their processing language (Albert & Obler, 1978) especially for children, ten years or
older, who are fully bilingual (Genessee, 1978). Stark, Genesee, Lambert and Seitz (1977)
determined that bilinguals developed more laterized verbal processing systems (Homel & Palij,
1987, p.68). Furthermore, bilingualism increases children’s mental flexibility and abstract
reasoning, facilitating their inclination to reorganize cognitive schemes (Ben-Zeev, 1977; Peal &
Lambert, 1962; Petitto, 2003), and promoting divergent thinking (Cummins, 1975).
The most frequently cited study comparing the cognitive performances of monolingual and
bilingual (French-English) students was conducted by Peal and Lambert (1962). They compared
English Speaking Canadians who attended French Immersion Public School in Montreal, Canada
with English Speaking Canadians who only received one period of French a day. Their study was
conducted during emotionally and politically charged times on all sides of the “language” debate,
before French was made the official language of Quebec. They suggested that two languages
result in “mental flexibility, a superiority in concept formation, and a more diversified set of mental
abilities in the sense that the patterns of abilities developed by bilinguals were more diversified”
(Lambert, 1972, p. 152; Lindholm, 1992, p. 206). Peal & Lambert (1962) concluded that bilinguals
performed significantly better than monolinguals on most measures of verbal and non-verbal
intelligence, tasks that required mental manipulation and reorganization of visual patterns, concept
formation, and tasks that required mental and symbolic flexibility (Lambert, 1972, p. 154).
Although Ben-Zeev (1977) makes the case that bilingualism retards vocabulary development, she
agrees that bilinguals devise ways to cope with the complexity of managing two languages,
therefore enhance their cognitive abilities. Ben Zeev argues that in organizing two linguistic
systems, bilinguals develop more analytic approaches, promoting cognitive growth (Been-Zeev,
1977, p. 1010).
After conducting the St. Lambert’s Experiment Study in Eastern Canada, Lambert (1972) suggest
that bilingual children perform better than monolingual children in cognitive tasks due to their
enriched experiences in two languages.
The emergence of an intellectual factor is dependent on the accumulation of experiences… The
bilingual child has been exposed to a wider range of experiences that the monolingual, because his
experience stems from two different [languages and] cultures. This enriched environment may
benefit him on nonverbal tests Higher levels of abstract cognitive processing are developed when
children learns to use two languages, generating symbolic reorganization, agility in concept
formation and improved performed at field independent tasks (pp. 121-40).
According to Baker (1988) the St. Lambert Experiment research project provides some
confirmation that bilingualism is linked to greater field independence. To the contrary, Genesee,
Hammers, Lambert, Monomen, Seitz & Starck (1978) claim that field independence tends to
preclude the acquisition of a second language. In addition, Cummins (1975) found that bilinguals
were superior at tasks that required divergent thinking.
Other frequently declared advantages to bilingualism are greater sensitivity to non-verbal
communications (Skutnaab-Kungas, 1981) and metalinguistic awareness (Bialyosky 1987,1988;
Hakuta &Diaz; 1985; Galambos & Hakuta, 1988; Lasagabaster, 1988; Mohanti, 1978, 1987;
Ricciardelli, 1992, 1993). Hamers and Blanc (2000) emphasize how bilinguals develop special
‘reflective’ skills, which foster other meta-cognitive processes contributing to bilinguals enhanced
learning skills (p. 91). These reflective skills may influence how children perceive and visually
represent events and emotions in their lives (Varon, 2000). Furthermore, Baker mentions studies
where “bilinguals pay attention to the fine details of each social situation” (Baker, 1988, p. 35).
Baker (1988) for example, in reviewing Bain (1975), states that bilinguals had greater sensitivity to
emotional expressions: they were able to better recognize appropriate feelings (Baker, 1988, p.
35).
Recent comparative studies conducted in the United States show how significant the educational
model and levels of bilingual proficiency is to the results. In the right context, two-way immersion
programs for the most part, bilinguals have demonstrated greater academic performances and
more positive cross-cultural attitudes (Cazabon, Lambert & Hall, 1993; Cazabon, Nicoladis, &
Lambert, 1998; Collier, 1987, 1995; Collier & Thomas, 2000; Genesee, 1999; Griego-Jones, 1995;
Lambert, 1977; Lambert & Tucker, 1972; Lambert & Gardner, 1973; Lambert, 1987; Lambert &
Cazabon, 1994; Hakuta, 1986, 1987, 1990, 1999; Skutnabb-Kangas, 1981; Thomas & Collier,
1997, 2002).
The parallel development of two languages is called “additive bilingualism” (Lambert, 1987) as
opposed to “subtractive” bilingualism, in which immigrants learn the language of the dominant
culture at the expense of their heritage language. Gonzalez (1999) affirms “additive bilingualism in
early childhood can accelerate cognitive growth if the child has relatively high proficiency in both
languages” (p. 22). Lindholm (1992) explains, “…additive bilingualism is associated with high levels
of proficiency in two languages, adequate self-esteem, and positive cross cultural attitudes” (p.
197).
Lambert contrasts the subtractive bilingualism of many minority language children with the additive
bilingualism generally achieved by children whose L1 [primary language] is dominant and
prestigious and in who danger of replacement by L2 [second language]. This is the situation of
Anglophone children in French immersion programs… The majority of studies reporting cognitive
advantages associated with bilingualism have been attained an additive form of bilingualism, that is,
relatively high levels of proficiency in both languages (Homel & Palij, 1987, p. 59).
Bilingual Education in the United States
In the US, the Bilingual Education Reform has inspired the implementation of maintenance and two
way immersion bilingual programs, providing more effective than other program models such as
structured immersion or transitional bilingual (TBE). In spite of current research findings pointing
to the most effective ways to educate bilingual children, the debate on the effectiveness of bilingual
education continues to be in the forefront of the political domain. In 1971, the Transitional
Bilingual Education Law was passed making bilingual education mandatory in Massachusetts. Any
district, where there are 20 or more Limited English speaking students from the same linguistic
background, the school system is required to offer native language instruction to those students
for up to three years.
Bilingual education began as a means of helping immigrant students understand school-work while
acquiring English, and maintain their linguistic and cultural identities. The law sought to provide
Limited English Proficient (LEP) students with greater opportunity for academic success than those
students who are immersed in English-only programs (Chapter 71 A, Transitional Bilingual
Education Law). As opposed to maintenance and two-way immersion programs, both TBE and
structured immersion do not target the development of students’ native language. Rather, they
promote the fast assimilation of students into all English programs, which in effect, devalue the
students’ native language.
Indeed, several reports (Porter, 1996; Rossell & Baker, 1996) have been critical of the bilingual
educational model called “Transitional Bilingual Programs”. This type of program was designed to
provide native language support while immigrant children learned English. In this educational
model, students gradually withdraw from all native language instruction, showing signs of what
Lambert (1972) called subtractive bilingualism. These critical reports subsequently do not refer to
detrimental effects of bilingualism as a condition, but rather attack the way children are treated in
the US when they enter transitional bilingual programs. “Teaching a LEP [Limited English
Proficient] student in the native language is at least marginally detrimental to his or her overall
education and acquisition of English” (Rossell & Baker, 1996, p. 2). These authors claim that
students enrolled in transitional bilingual programs will consequently fall behind in all academic
subjects.
In California (Proposition 227, 1999), Arizona (Proposition 203, 2000) and very recently
Massachusetts (English for Children Initiative), bilingual education has already been abolished
through state referendums and replaced with structured immersion programs. One exception, so
far, is in Colorado where the citizens voted to protect their bilingual programs. In spite of the
political trends, research on bilingual education suggests that two-way bilingual immersion
education is effective in educating balanced bilingual children because it is based on an additive
model of bilingualism. In the two-way immersion model of implementation, students from diverse
linguistic backgrounds develop and master the content areas in two languages simultaneously.
Two-Way Bilingual Immersion Programs
Two-way or dual-immersion programs were based on the Canadian model of immersion programs
for foreign language instruction, yet they also integrate elements of the developmental bilingual
model for native language instruction. In the United States dual-immersion programs have
provided equal opportunity for Limited English Proficiency (LEP) students and Native English
Speakers from diverse linguistic backgrounds to be integrated for all subjects and each group to
become proficient bilinguals.
Also during the last three decades, the dual immersion bilingual context has been fertile ground for
research studies, and consensus exists as to their benefits in academic, social, and cognitive areas.
In fact, students in two-way programs have been shown to develop positive cross-cultural
attitudes and they out-perform their monolingual counterparts in academic subjects such as
language arts and mathematics (Cazabon, Lambert & Hall, 1993; Cazabon, Nicoladis, & Lambert,
1998; Christian, 1994; Collier & Thomas, 2000; Genessee, 1987, 1994, 1999; Howard, 2000;
Lambert & Cazabon, 1994; Lindholm, 1992, 1992; McLaughlin, 1995; Thomas & Collier, 1997,
2002).
Collier and Thomas (2002) conducted a five-year longitudinal study that examined academic
achievement of 210,054 students (80 primary languages) from eight different models of bilingual
program including two-way immersion. The demographics in three of the five research sites
focused on Spanish speakers, the largest group of school age population (75% of US linguistic
minority students). The study used standardized tests used by the five urban and rural research
sites across the United States. In general, Collier & Thomas found that bilingual students out-
performed their monolingual counterparts in all academic subjects at the elementary level.
Bilingually school students outperform comparable monolingually school students in academic
achievement in all subjects, after 4 to 7 years of dual language schooling (p. 9).
Native English speakers in two-way bilingual immersion programs maintained their English, added a
second language to their knowledge base and achieved well above the 50th percentile in all
subjects areas on norm-referenced tests in English (p.7).
Thomas and Collier found that Native Spanish Speakers also benefited from the two-way
immersion program model, achieving a 71st percentile after two years in the US.
As of today, no studies investigating the effects of bilingualism have employed the analysis of
students’ drawings to examine cognitive outcomes. Since drawings manifest students’ abstract
thinking as well as their internal representations of social behaviors, and “during the elementary
school years, boys and girls can express their thoughts and feelings often better in visual images
than in words” (Koppits, 1984 p. 2), for my dissertation, I will analyze drawings to examine
students’ cognitive aptitudes.
Hypothesis
I anticipated that bilingual speakers (ages 10-12) would reflect higher levels of cognitive complexity
than those produced by their monolingual counterparts in their graphic compositions, visual
representations of interactions and their verbal descriptions of portrayed interactions.
I anticipated that bilingual speakers (ages 10-12) would reflect higher levels of cognitive complexity
than those produced by their monolingual counterparts in their graphic compositions, visual
representations of interactions and their verbal descriptions of portrayed interactions. My results
may coincide with Vygotsky (1962) in his idea that knowing two language-systems influences the
cognitive condition, suggesting that enhancing development of one domain determines
development in another. This view considers that social and cultural influences (language in
particular) determine and precede cognitive capacities. This view considers that teaching propels
or accelerates development; thus development is context specific.
Vygostky (1935, p.14) suggested that when the application of sound pedagogical principles
ensured that each language had an independent sphere of influence, bilingualism could orient the
child towards more abstract thought processes “from the prison of concrete language forms and
phenomena (Cummins in Homel & Pajil, 1987, pp. 58-59).
Conversely, my results may support Piaget’s claim that cognitive development is concurrent or
precedes linguistic development. Both cognitive and linguistic developments are linked to the early
development of intelligence and to a child’s capacity for symbolic representation (mental
structure). This view postulates that development is universal and not culture specific and that it is
achieved through the manipulation of objects (not necessarily through the development of
language).
Piaget argues that these stages in cognitive development are neither biologically programmed nor
given by the environment, but instead constructed or created by children as they actively attempt
to understand the world (Feldman, 1999, p. 59).
In short, even though both Vygotsky and Piaget agree that environmental influences determine
children’s cognitive development, Piaget attributes development to the interaction between a child’
s innate capacity and his or her experience (assimilation and accommodation), whereas Vygotsky
considers that language exists in thought and that with systematic cooperative interactions
development can be produced regardless of a child’s innate condition.
Both monolingualism and bilingualism are linguistic conditions. One language may exist separate
and may function as a concrete, linear, successive, and consecutive system. Two languages may
exist simultaneously and are always in relation to one another, possibly producing additional
reflections, connections and abstractions, which in turn may function as concurrent and
synchronized abstract systems. Thus, my overarching question: How do these possible different
linguistic processes (consecutive vs. concurrent) impact children’s cognitive structures? If so, how
are they captured in children’s drawings?
In fact, Been Zeev (1977) found that the precocious appearance of categorical associations in
bilingual children developed earlier orientation towards “reorganization”. Bilinguals performed
better at “symbol substitution” tasks, developed an increased ability to analyze non-verbal
structures, and exhibited more mature behaviors than their monolingual counterparts, which
suggests bilingualism promotes precocious development (p. 1015).
When individuals become bilingual, they develop the ability to go from one language to the other,
transfer language skills from one language to the other, and arrive at levels of abstraction that
monolingual individuals are less likely to.
Along the same lines, Lambert (1987) and Bialystock (1999) suggest that bilingual children develop
“some sort of a tri-dimensional view of language, a stereo-scopic perception to which the
monolingual child does not access” (Hamers & Blanc, 2000, p. 85). This relates to Petitto’s finding
that the learning of two languages, in spite of their ability to switch from one language to the
other, does not confuse bilingual children.
The ability to know and to use two or more languages fluently and appropriately reflects the
remarkable functioning of a cognitive system that maintains separation among languages being
used while at the same time allows a free and easy interchange among them (Homel and Pajil 1987,
p. 131).
Research Questions
1. How do the cognitive structures of monolingual and bilingual speakers compare when
performing in three distinct modes of representation: graphic composition, visual content, and
verbal narrative?
2. How do the cognitive structures of monolingual and bilingual speakers compare when
examining their representations of three distinct emotions (happy, sad and angry)?
3. How do monolingual and bilingual speakers (grouped in three distinct ways) compare in
degrees of cognitive complexity as manifested in the graphic composition of their drawings?
4) How do they compare in the degrees of cognitive complexity manifested in the visual
representation of interaction in their drawings?
5) How do they compare in the degrees of cognitive complexity manifested through the verbal
descriptions of depicted interaction as measured by skills theory?
Methodology
Setting
The study was conducted in an urban school (K-8) in Massachusetts with approximately 550
students, 61 % of which participated in a free or reduced lunch program. The setting was ideal
because two different instructional programs, monolingual and two-way bilingual immersion, were
housed in one building. There were two bilingual classrooms and one monolingual classroom in
each of the participating fifth and sixth grades.
Subjects
Subjects attended the same public school in Cambridge, MA. Monolingual (42.39%) and bilingual
(57.61%) speakers were enrolled (for the most part) in two distinct educational programs:
monolingual (English only) program and a bilingual (Spanish/English) dual immersion program.
There were 9 Spanish monolingual students in the bilingual setting who were recent immigrants,
and 9 bilingual (Spanish/English or Portuguese/English) students enrolled in the monolingual
program whose home-language is other than English.
Data was collected from a total of 92 students, 39 of which were enrolled in the monolingual
program and 53 in the bilingual program. Forty-two were girls (45.65%) and 50 were boys
(54.65%). Half of the students (50.55%) participating in the study were eligible for free lunch, 5
students (5.49%) received reduced lunch, and 40 students (43.96%) pay for lunch. Half of the
students (50%) were Hispanic, a third were Caucasian (32.61%) and the rest were African
American (14.13%), Native American (2.17), and Asian (1.9%).
My final sample size is not large, which leads me to recognize that my findings are not generalized
to student-populations at large. I was, however, able to carry out a number of convincing analyses
that reveal hints or trends and that contribute to our understanding of the development of skills
among monolingual and bilingual students as I carried out an in-depth investigation of three
drawings per student. In all, 92 students produced 276 drawings.
Raven Progressive Matrices Test
To insure that groups (monolingual and bilingual) are similar in their basic levels of intelligence, I
obtained scores from their performance on the Raven Progressive Matrices (Raven, 1986) and
groups were matched. An expert who administers and scores this test for the school system
provided me with students’ scores. I used their scores in my analyses to be sure that groups are
not different in basic level of intelligence.
Non-verbal measures of intelligence assure that groups are similar in kind.
In order to isolate the effects of linguality, bilingual and unilingual groups should be matched on
any personal or background characteristic which might contribute to performance on the
dependent variables (Cummins, 2001, p. 33).
The Raven Progressive Matrices have been widely used as a non-verbal measure of intelligence.
Before Lambert’s St. Lambert Experiment (Lambert, 1972), The National Society for the Study of
Education (1928; 1940) published a series of studies that measure the impact of bilingualism on
intelligence citing positive effects (Goodenough, 1928, 1949; Terman 1928, 1949; Stoddard &
Wellman, 1928, 1949).
Distribution of Students Participating in the Study
The breakdown of students by grade level indicates that in fifth grade (ages 10-11) there are a
total of 42 students: (18 (19.57%) in monolingual classes and 24 (26.09%) in bilingual classes).
In the sixth grade (ages 11-13) there are a total of 50 students: 21 students (22.83%) in
monolingual and 29 students (31.52%) in bilingual classes).
Table 1
Number of Students per grade and per educational program
Students 5th Grade 6th grade Total
Monolingual Program 18 21 39
Students in Two-Way Bilingual Program 24 29 53
Total 42 50 92
When grouping students according to their linguistic abilities, there were 53 (57.61%) bilingual
speakers and 39 (42.39%) monolingual speakers.
Table 1.1
Distribution of students according to their linguistic ability
Grade English Monolingual in Monolingual
Program (MM) Spanish Monolingual in Two-Way Bilingual Program (MB) Bilingual in
Two-Way Bilingual Program (BB) Bilingual in Monolingual Program (BM)
5th 12 (13.04%) 4 (4.35%) 20 (21.74%) 6 (6.52%)
6th 18 (19.57%) 5 (5.44%) 24 (26%) 3 (3.26%)
Total 30 (32.61%) 9 (9.78%) 44 (47.83%) 9 (9.78%)
39 (42.39%) 53 (57.61%)
Procedures
Data Collection
I collected three drawings per student in three consecutive sessions, from a total of 92 students
with 2 different linguistic abilities, in either a monolingual or a two-way bilingual immersion
programs. During each session, students worked in their classroom settings and produced one
drawing per session. For practical reasons, since there were two bilingual classrooms per
monolingual one for each grade level. I grouped students from bilingual classrooms of the same
grade level for data collection purposes. Students worked individually, sitting in small groups, while
classroom-teachers were present. Introductory prompts were given orally and most always in
English. If needed, students translated for each other.
All data collection sessions and follow up interviews were observed for bias by a constant outside
witness whose observations were done according to protocol (Appendix B). This individual
was unfamiliar with the school and the distinct programs where data was collected.
To scaffold students’ efforts (Vygotsky, 1978), I showed book-illustrations of relationships and
specific emotions (happy, sad or angry) and encouraged students to share what they saw. All
students looked at the same illustrations before I introduced the task to the whole class at once.
Students each drew a picture after I read the “happy” prompt the first session; “sad” prompt the
second session and “angry” the third session (Appendix C). The logic used for ordering the tasks
was that “happy” was best to start with, to set the tone; “sad” was second because I did not want
it to be last; and “angry” was third by default.
I provided students with pencils, paper, and asked them to indicate their names, grade level, and
classroom number on the back of their papers. Children worked until they considered their
drawings finished, which typically took approximately 20 minutes. Students held on to their
drawings and waited to be called for a follow-up interview.
Interviews
Immediately following each drawing session, I conducted brief individual interviews (approximately
3-10 minutes per student) in an adjacent library. Children’s descriptions of their drawings using
questions (Appendix D) designed to capture their intentions, thoughts, and descriptions of the
relationships, places, and interactions portrayed in each drawing were tape-recorded. I
encouraged students to give as much information about their drawings as they could.
To ensure the most eloquent narratives all Spanish monolingual students in the bilingual program
were interviewed in Spanish; and all bilingual students in the monolingual classrooms, as well as all
other participants, were interviewed in English.
Scoring Procedures
Inter-rater Reliability.
An independent rater and I scored 20 % of all drawings for their graphic composition, 20 % of all
transcriptions for their verbal narrative, and 40 % of all drawings for their visual content. This was
done to determine if different raters could reliably use the scales.
The independent rater for the drawings was a veteran art teacher with a Bachelor’s in Fine Arts
and Psychology from Emmanuel College and a Masters Degree in Expressive Therapies from Lesley
University.
The independent rater for the verbal narratives was former student of Kurt Fischer who graduated
with a Master’s Degree from Harvard University. For the verbal accounts of the nature of
interaction in students’ drawings, ratings were done based on the content in students’ answers
(from transcribed tapes) using Fischer’s criteria for the various levels of complexity (Appendix A).
Both raters demonstrated great understanding of the theories and scales that they used and had
no previous knowledge of my study or the students who participated in it. During the process,
raters had no indication of students’ linguistic classification, grade level, or program enrollment.
For practical purposes, students were assigned identification numbers and data sets and scoring
spreadsheets were color-coded.
The levels of agreement between raters were very high:
- for the graphic composition we agreed on 48 out of 53 drawings = 91%
- for the verbal narratives we agreed on 50 out of 53 drawings = 94 %
- for the visual content we agreed on 81 out of 91 = 89%
These results indicated that the independent raters could continue finishing the rest of the scoring
on their own.
Scoring of Drawings by Independent Raters.
Each student received three scores per drawing, one in each category: graphic composition, verbal
description of interaction, and visual representation (a total of nine scores for three drawings –
three for each category). Values were pre-designated for each category (Table 2) and criteria for
each category was determined
Table 2 –
Values for Scoring Drawings in Three Categories
Category 1 Category 2 Category 3
GRAPHIC……………….. VERBAL …………………VISUAL
Based on Piaget Based on Fischer Proposed by
Varon
and Lowenfeld (1947) (1980) (2002)
Golomb (1992) Di Lio (1977)
Scores were recorded on color-coded scoring sheets. The scales for graphic composition and
verbal description of interactions had been widely used to measure artistic and cognitive
development. The scale for visual representation was derived from Fischer (1980) and was my
proposed measure to identify cognitive development through drawings. Values in each category
were linked to specific criteria (Appendix A) representing progressive levels of cognitive
development (from simple to complex) from three different perspectives:
1) Graphic composition based on traditional artistic development by looking at the
arrangement and perspective of visual items on the page as identified by Lowenfeld (1939). The
scale (Table 2.1) consisted of numbers from 1 to 4, with 4 capturing higher levels of graphic
composition.
Table 2.1 –
Criteria for Scoring the Graphic Composition Category:
STAGE DESCRIPTION VALUE
Pre-Schematic:
1. Simple: Floating figures or attached to a base line; single point of view with action related to the
self only.
Random collection of objects or figures that do not seem to relate to each other on the page;
- No depth or overlap;
- Eye level view: without any horizon line; or
- Single line
or
- Perpendicular Lines relationships are used to depict space:
- Side or Front view of same object limited to two faces of object or agent, though front face
preferred;
- Transparencies or line Intersections are present 1
Visual Realism
2. Multiple; Figures relating to more than one linear axis. Objects or individuals are arranged
representing the personal perspective of self.
- Children are able to use more realistic proportions of figures and portray them in profile
- Two separate ground lines or combining birds eye view
- Children use fold over perspective - overlaps
- Story line and time sequences are depicted in representation
2
Realism /Literal
3. Sequential: sequence of time and space defined in frames or boxes creating a narrative yet still
from the perspective of self.
- Children are able to depict differentiation between self and others each with own characteristics –
true to appearance;
- Cross gender characteristics
- Detail oriented drawing
- Children are able to depict figures in motion
- Figures are represented only part view (Lange-Kuttner & Wolfgang Edelstein p.71)
- Cartoon-like at times
- Parallelogram perspectives is used - frontal parallel plane 3
Pseudo-Naturalism
4. Global: perspective determined by global composition, combining different points of view, not
limited to the perspective of self.
Children are able to draw the idealized human form
Children are able to use converging perspectives: combining vertical and oblique projections –
Children pay attention to mood and often representations are personal statements.
Children pay attention to greater detail and or abstractions depicting cause and effect/
interdependence of relations.
Children are able to represent three dimensions 4
Coordination of Abstractions
Children are able to create groups of concepts
Simple symbol relationship
2) Verbal description of depicted interaction by looking at the nature of roles and relationships
as evidence of the complexity level as postulated by Fischer (1980, 1993). The scale here ran from
1 to 5, with 5 representing higher levels of complexity (Table 2.2).
3) Visual representation of self in relationship with others by examining the nature of
portrayed relationships and objects in the page rather than composition of graphic elements of
drawings. Levels (Table 2.2) were inferred from Fischer’s cognitive levels and adapted to create an
alternate way of classifying artistic development. This will be scored using the same 1-5 scale of
complexity.
Table 2.2 - Criteria for Scoring the Verbal & Visual Representation Category
Fischer’s Levels – Verbal Relations Score Varon’s Levels - Visual
I- Single Representations
(Rp1)
(Descriptions of criteria provided by Fischer 2003 in e-mail correspondence) No relation
1 Discrete – Distinct Image
Child draws one or more figures or objects without evidence of any relationships or specific location
II- Representational Mappings
(Rp 2)
Involves mapping of two or several items focusing on one aspect of each
Reciprocal
Sequential
2 Coordinated – Sequential Image
Child draws two or more figures/objects in a direct, simple relationship either collaborative or
complementary. Arrangement of figures and objects on the page are not unified or
interconnected.
Child draws figures involved in the same/similar action with little or no differentiation between
them. Child draws figures or object in a generalized context.
III Representational Systems
(Rp 3) Involves relating several aspects of each item in a system Causal
Coordinated
Temporal
Instrumental 3 Integrated Image
Child draws two or more figures/objects that are arranged in an integrated system of concrete
relationships, - At times figures or objects are portrayed in synchronized action/s that produce
reactions within that system. The cause, reason, and context (specific place) for actions are explicit.
Child draws evidence of a temporal relationship.
IV– Single Abstractions
(Rp 4 or Ab 1)
Involves coordinating several concrete inter-relationships (systems) in terms of an abstract or
intangible characteristic
Intangible
Conflictive 4 Figurative – Symbolic image
Child draws one or more concrete figures/objects forming relationships that produce intangible
meaning. These relationships have the clear intent to express implicit characteristic.
Child draws one or more figures/objects involved in differentiated systems of relationships that
result in the portrayal of conflictive or distinct values or attributes.
Child uses universal symbols to express abstract concepts.
V – Abstract Mapping
(Ab 2)
Involves relating two abstracts characteristic with each other Deduction
5 Logical Symbolic image
Child draws one or more figures/objects that are portrayed through symbols expressing a desired
outcome/idea. The image carries clear motivation in the use of symbols - or symbolic objects - to
depict what self intends.
Child combines sets of symbols to convey a meaning beyond the symbols themselves.
For all scoring, two individuals (an independent rater and me) scored and coded at least 20 % of
the data set separately. Where differences in scores appeared between raters, discussions were
held until inter-rater agreement was reached. The independent rater provided the scores used for
the final analysis.
Analysis
To examine monolingual and bilingual students’ drawings, I designed an integrated task that
included three components of what constitutes a representation according to Freeman (1980): a
definable object, definition of a relationship, and the study of action. In the analysis I looked at the
mean scores of group performances in each of three categories (graphic, verbal, and visual) and
translated them into percentages because the scales used to assign scores in each category were
not the same. In order to find the corresponding percent for each category, I divided the score by
the number of steps on each scale, (i.e. all scores for the graphic composition were divided by four;
all scores for the visual and verbal categories were divided by five). Once scores were comparable,
I examined the data by grouping students in three different configurations looking for changes in
their representational structures and the organization of the represented relationships. The
grouping of students followed three different configurations:
1) Monolingual and bilingual speakers;
2) Four distinct sub-groups of monolingual and bilingual speakers grouped according to their
linguistic backgrounds and educational experiences; and
3) Students enrolled in two distinct educational programs (English Monolingual and Two-Way
Spanish/English Bilingual Immersion).
For each configuration of students, I organized the data according to:
a) Three representational modes: graphic, visual, and verbal; and
b) Three distinct emotions: happy, sad, and angry.
Answering Research Questions
My main research questions focuses on the differences in the degrees of cognitive complexity
between the monolingual and bilingual speakers and the differences between students enrolled in a
monolingual English program and a two-way bilingual immersion program. I analyzed three modes
of representation and the representation of three emotions.
1.) As reflected by the three modes of representation together;
2.) As reflected by the representation of three emotions together;
3.) As reflected in the graphic composition of students’ drawings;
4.) As represented in the relationships of the visual content depicted in students’ drawings;
and
5.) As reflected in students’ verbal descriptions of the interactions in their drawings.
I addressed the research questions by carrying out a series of regression analyses, using separate
analyses for each of the main outcomes. The scores from three drawings and corresponding verbal
narratives (for happy, sad, and angry) were analyzed separately and combined. I also combined
the scores from the three different modes of representation: graphic, visual, and verbal. The main
question predictor variable was bilingual status (coded 0 = monolingual program / 1= bilingual
program); and the other predictors were student linguistic abilities (coded 0= English-only / 1 =
Other than English (mostly Spanish or Portuguese + English); grade level (coded 0 = fifth grade /
1= sixth grade).
I examined if the effects of bilingual status or program enrollment would be significant to either the
mode of expression and the emotional content of students’ representations. Thus, I investigated
possible interactions of linguistic and educational status, home language, and emotional content of
drawing. Significant interaction effects would indicate that certain subgroups of students
(Appendix F) perform differently than others.
Finally, I examined the differences between the three scales (different ways to score drawings) and
the relationships between the emotions represented in each drawing. This helped me to see
whether students are relatively consistent in their scores on graphic compositions, visual
representations, and verbal descriptions of interaction, or whether there are notable differences
across the emotions.
Findings
Overall, there were no significant statistical differences across three of the four compared groups.
It is interesting to note that the Spanish monolingual group who scored statistically lower in the
Raven Progressive Matrices showed the few instances where there was a statistical difference.
Comparison Of Students’ Basic Level Of Intelligence
In comparison of the basic level of intelligence using the Raven Progressive Matrices the differences
were statistically significant in all three comparisons: a) across four groups of students, (F=7.71,
p= 0.0067); b) across monolingual and bilingual speakers (F=4.31, p=0.04); and c) across
educational programs (F=4.04, p=0.047).
Pair-wise comparisons indicated that there was a consistent difference between the Spanish
monolingual speakers in the bilingual program (one sub-group) and everyone else. The Spanish
monolingual speakers in the bilingual program scored the lowest (M=45.55%, SD = 36.43) of the
four sub groups. Their score is statistically significantly different than the bilingual speakers in the
bilingual programs (p=0.0007), than the bilingual speakers in the monolingual program (p=0.
0246), and than the monolingual speakers in the monolingual program (p=0.0013). (See table 3.)
Once I segregate the scores of the Spanish monolingual students, all other comparisons of the
Raven scores do not reveal significant statistical differences. Bilingual speakers in the bilingual
program (M=8.25, SD = 29.50) and English monolingual speakers in the monolingual program (M=
8.2, SD= 27.46) scored practically the same with no statistically significant difference between
them (p= 0.9418). Bilingual speakers in the monolingual program (M=76.66%, SD =
20.61) scored in third place, also with no statistically significant difference between them and
bilingual speakers in the bilingual program (p= 0.5821) and the English monolingual speakers in the
monolingual program (p= 0.6281).
The Spanish monolingual students are, for the most part, recent immigrants to the United States
who did not receive comparable contextual and environmental influences while growing up
(Appendix F). Results are presented in table 3.
Table 3
P-Values when Comparing Raven Progressive Matrices Across Four Groups
Bilingual Speakers In Bilingual Program
(BB) Bilingual Speakers In Monolingual Program
(BM) Spanish Monolingual Speakers In
Bilingual Program (MB) English Monolingual Speakers in Monolingual Program (MM)
BB - 0.0007 0.5821 0.9418
BM - - 0.0246 0.0013
MB - - - 0.6281
MM - - - -
The above-mentioned differences between the sub-group of Spanish monolingual students and
the rest affect all comparisons in the various configurations. When comparing the Raven scores of
students in the monolingual (M=80.76%, SD = 25.89) and bilingual (M=76.22%, SD = 33.46)
programs, the difference is statistically significant (F=4.04, p=0.0475). In addition, the difference
in Raven scores between the monolingual (M=73.58%, SD = 33.12) and bilingual (M=81.50%,
SD= 28.10) speakers, which also includes the sub-groups, is statistically significant (F=4.31. p=0.
0408).
A - Aggregated Scores for Three Expressive Modes of Representation and the Representations of
Three Emotions
When aggregating scores of all students and examining the differences across the three modes of
representation (Table 3.1 and Appendix G), the contrast shows that graphic composition
generated the highest score (M=57.78%, SD=15.11), followed by visual representation (M=47.
68%, SD=12.31), and then by verbal narrative (M=39.92%, SD=11.06).
There were overall differences in the levels of complexity ratings across the three modes (F (2,182)
= 73.1, p= <0001). Results are shown in table 3.1. The post-hoc contrasts here revealed that the
ratings for graphic were higher than those for verbal (p=.0001) and visual (p-.0001) and the
difference between verbal and visual was also significant (p=.0001).
When aggregating scores of students in their representations of events associated with three
emotions (Table 3.1 and Appendix H) and examining their differences across emotions, “anger”
produced the highest score (M=50.34%, SD=14.28) which reflects the highest level of complexity,
followed by “sad” (M=49.64%, SD=14.40) in second place and “happy” which got the lowest score
(M=45.29%, SD=11.76).
When looking at the overall performance of all students in their levels of complexity across the
three different emotions (happy, sad and angry), I found that there was a statistically significant
difference (F (2,178)= 5.62, p = 0.004). Results are shown in table 3.1. The post-hoc contrast
indicated that the mean score for happy was significantly lower than the means for sad (p = 0.004)
and angry (p = 0.016). The means for sad and angry were not statistically different.
It seems interesting to note that I found a strong correlation between the visual and graphic
modes (r = .81) and weak correlations between the visual and verbal modes (r = .18); and the
verbal and graphic modes (r = .21). This indicates that it is important to assess children using
different modes if we expected to construct a fuller picture of their cognitive abilities.
This is evidence that the Varon scales, derived from Fischer’s scales for measuring cognitive levels
through behavior and thinking skills, are successful in measuring cognitive development through
drawing. The Varon scales were therefore developmentally appropriate.
The verbal mode and happy emotions produced the lowest levels of complexity, contrary to
graphic representations and anger, which reached the highest levels of complexity. Results are
presented in table 3.1.
Table 3.1
Average Ratings found in Modes of Representation and Representations of Emotions
Level of Complexity Modes of Representation
(F (2,182) = 73.1, p= <0001) Representation of Emotion
(F (2,178)=5.62, p =0.004)
Highest levels Graphic: M= 57.78%, SD=15.71 Angry: M= 50.34%, SD=14.28
Middle levels Visual: M= 47.68%, SD=12.31 Sad: M= 49.64%, SD=14.40
Lowest levels Verbal: M= 39.92%, SD=11.06 Happy: M= 45.29%, SD=11.76
B - Comparison of Monolingual and Bilingual Speakers’ Performance In Three Representational
Modes And Three Emotions (See Table 4 & Appendix I)
When comparing Monolingual and Bilingual Speakers’ aggregated scores in the three modes of
representational systems, F-Tests results indicate that there is no statistical difference between the
two groups (F=1.41, p=0.2381). Bilingual speakers scored only very slightly higher (49.17%, SD=
9.63) than monolingual speakers (47.50%, SD=11.21).
Additionally, when comparing monolingual and bilingual speakers’ aggregated scores in the three
representations of emotions, F-Tests results indicate that there is no statistically significant
difference between the two groups (F=1.43, p=0.2356). Bilingual speakers scored slightly higher
(49.12 %, SD= 9.64) than monolingual speakers (47.48%, SD=11.22).
Table 4
Monolingual and Bilingual Speakers’ Average Rating in all Three Modes of Representation and
Three Emotions
Monolingual Speakers
N = 39 Bilingual Speakers
N = 53 ANOVA
Variables M SD M SD F p
Modes of Representation 47.50 % 11.21 49.17% 9.63 1.41 0.2381
Representation of Three Emotions 47.48% 11.22 49.12% 9.64 1.43
0.2356
B.1- Monolingual And Bilingual Speakers’ Performance In Three Modes Of Representation (See
Table 4.1 & Appendix J)
Results in the three distinct representational modes indicate that there is no significant statistical
difference found between monolingual and bilingual speakers.
Graphic Composition. F-Tests results indicate that there is no statistical difference between the
two groups (F=2.29, p=0.1334) when looking at the average scores of monolingual and bilingual
speakers based on their graphic composition (in a scale of 1-4 based on Piaget and Lowenfeld’s
levels of artistic development), bilingual speakers (58.33%, SD=14.52) scored slightly higher
(57.05%, SD=17.36) than monolingual speakers in graphic representations.
Visual Representation. F-Tests results indicate that there is no statistical difference between the
two groups (F=2.84, p=0.0956) when looking at the average scores of students’ drawings based
on the complexity of their visual representations, (in a scale of 1-5 based on Varon’s adaptation of
Kurt Fischer’s levels of cognitive development), bilingual speakers scored higher (M= 49.55%,
SD=11.74) than monolingual speakers (M= 45.12%, SD= 12.74) in the representation of visual
content.
Verbal Narrative. F-Tests results indicate that there is no statistical difference between the two
groups (F=0.44, p=0.5071) when looking at the average scores of students’ verbal narratives (in a
scale of 1-5 based on Kurt Fischer’s levels of cognitive development). Monolingual speakers (M=40.
34%, SD=12.60) scored slightly higher than bilingual speakers (M=39.62%, SD=9.88) in their
verbal narratives.
Table 4.1
Monolingual and Bilingual Speakers’ Average Ratings in Three Representational Modes
Monolingual Speakers
N = 39 Bilingual Speakers
N = 53 ANOVA
Variables M SD M SD F p
Graphic 57.05 17.36 58.33 14.52 2.29 0.1334
Visual 45.12 12.74 49.55 11.74 2.84 0.0956
Verbal 40.34 12.60 39.62 9.88 0.44 0.5071
Total 47.50 11.21 49.17 9.63 1.41 0.2381
B.2. - Monolingual And Bilingual Speakers’ Performance In The Representations Of Three Distinct
Emotions (See Table 4.2 & Appendix K)
Results in the three distinct representations of emotions indicate that there is no significant
statistical difference found between monolingual and bilingual speakers.
Happy. The F-Test results indicate that there is no statistically significant difference in the
mean (F=1.33, p=0.2524) in students’ representation of happy. Monolingual speakers (M=45.40
%, SD = 11.05) and bilingual speakers (M=45.22%, SD = 12.36) scored virtually the same.
Sad. The F-Test results indicate that there is no statistically significant difference in the mean
(F=1.13, p=0.2909) in students’ representation of sad. Bilingual speakers scored higher (M=51.
53%, SD=13.77) than monolingual speakers (M=47.17%, SD=15.01).
Angry. The F-Test results indicate that there is no statistically significant difference in the
mean (F=0.34, p=0.5617) in students’ representation of anger. Bilingual speakers scored higher
(M=50.69 %, SD=13.12) than monolingual speakers (M=48.87 %, SD=15.89).
Table 4.2-
Monolingual and Bilingual Speakers’ Average Ratings for the Representation of Three Emotions
Monolingual Speakers
N = 39 Bilingual Speakers
N= 53 ANOVA
Variables M SD M SD F p
Happy 45.40 11.05 45.22 12.36 1.33 0.2524
Sad 47.17 15.01 51.53 13.77 1.13 0.2909
Angry 49.87 15.89 50.69 13.12 0.34 0.5617
Total 47.48 11.22 49.12 9.64 1.43 0.2356
C. - Comparison of the Performances of Four Groups of Students Based On Their Linguistic
Backgrounds
This comparison examines the data of four distinct groups of students based on their linguistic
backgrounds and enrollment in educational programs:
(BB) Proficient bilingual students enrolled in bilingual two-way immersion program who
come from Bilingual, Spanish, or English Speaking homes (n = 44);
(MM) English monolingual students who are enrolled in a monolingual program who are from
English-only speaking homes (n=30);
(BM) Bilingual students whose home language is other than English who are enrolled in the
monolingual program (n=9), some of whom are international students; and
(MB) Spanish monolingual students who are native Spanish speakers who are enrolled in the
bilingual classrooms (n=9). These students were all recent arrivals to the United States and new to
the school system.
C.1. - Aggregated Scores For Three Representational Modes (See Table 5 & Appendix L)
The differences across four groups are not statistically significant (F=1.53, p=0.2191) when
looking at the average aggregated scores for all three representational modes expression
together. Bilingual students in two-way bilingual programs scored the highest (M= 49.24%, SD=
9.17). English monolingual students in monolingual program (M= 48.96%, SD=11.99) followed.
Bilingual students in monolingual program, ranked in third place (M= 48.82%, SD=12.28), and
Spanish monolingual students in two-way bilingual program scored the lowest (M= 42.65%, SD=
6.43).
These findings are consistent with differences found in the Raven Progressive Matrices
performances. In effect, even though the differences here are not statistically significant, it is
worth noting the difference in performance between Spanish monolingual students in the two-way
bilingual program and all other three groups. These Spanish monolingual students’ lower
performance is consistent with their lower performance in the Raven Progressive Matrices which is
most likely influenced by their underprivileged backgrounds, limited schooling, and restricted
chances to express themselves in two languages in school.
C.2. - Four-Group Performance In Three Distinct Modes of Representation (See Table 5 & Appendix
M)
When data was desegregated by different representational modes, one of the sub-groups
performed consistently lower than the other groups, but the difference was not always statistically
significant.
Graphic Composition. In graphic composition (in a scale of 1-4 based on Piaget and Lowenfeld’s
levels of artistic development), differences across groups were not statistically significant (F=1.03,
p=0.3121). The only statistically significant difference was when comparing the Spanish
monolingual speakers in the bilingual program as compared with the English monolingual (p=0.
0328) and bilingual (p=0.0457) speakers in the monolingual program. The difference is not
statistically significant when comparing the Spanish monolingual speakers in the bilingual program
with the bilingual speakers in the bilingual program (p= 0.0714).
In their graphic compositions, bilingual speakers in the monolingual program scored highest (M=62.
03%, SD=15.65), followed by English monolingual speakers in monolingual classes (M=60.00%,
SD=17.96). In third place were the bilingual speakers in the bilingual program (M=57.57%, SD=14.
35), and last were the Spanish monolingual speakers in the bilingual program (M=47.22%, SD=11.
02).
Visual Representation. There is no statistical difference across the groups (F=0.74, p=0.3926)
when looking at the average percent scores of students’ drawings based on the complexity of their
visual representations, (in a scale of 1-5 based on Varon’s adaptation of Kurt Fischer’s levels of
cognitive development). Bilingual speakers in the bilingual program scored highest (M=49.69%,
SD= 11.01), followed by bilingual speakers in the monolingual program (M=48.88%, SD=15.63),
then by English monolingual speakers in monolingual program (M=46.22%, SD=13.77). Spanish
monolingual students in bilingual program scored the lowest (M=41.48%, SD= 8.01). These
Spanish monolingual speakers in bilingual classrooms are recent immigrants from Spanish Speaking
countries or Puerto Rico and their “slight lag in development” may be due of diverse environmental
context and psycho-socio-economic factors affecting their educational backgrounds.
Verbal Narrative. There are no statistically significant differences across groups (F=1.15, p=0.
2863) when looking at the average scores of students’ verbal narratives (in a scale of 1-5 based on
Kurt Fischer’s levels of cognitive development). English monolingual speakers in monolingual
classrooms (M= 40.66%, SD=13.82) and bilingual speakers in bilingual classrooms (M= 40.45%,
SD=9.55) scored virtually the same in verbal narratives. Spanish monolingual speakers in the
bilingual programs (M=39.25%, SD=7.77) rated in third place (and most likely because their
interviews were held in Spanish) with a slightly better score than the bilingual speakers in
monolingual programs (M= 35.55%, SD=11.05).
The fact that the bilingual speakers in monolingual programs scored lower in verbal narratives than
the other groups, especially when we comparing their scores with the bilingual speakers in the
bilingual program, suggests that the two-way bilingual immersion program would be more a better
education fit for these students’ verbal development. As slight as this difference may be
(approximately 5 points), the impact of total assimilation and subtractive bilingualism prevents
bilingual home speakers from reaching their full cognitive potential. When students share and
preserve their native languages and cultures they seem to reach optimal potential. A study with
much larger number of students would be necessary to look for evidence of this hinted trend.
Table 5 -
Average Ratings of Drawings Across Four Groups and Across Three Modes of Representation
English Monolingual Speakers in Monolingual Program
(MM) Spanish Monolingual Speakers in Bilingual Program
(MB) Bilingual Speakers in Bilingual Program
(BB) Bilingual Speakers in Monolingual Program
(BM)
ANOVA
Variable M SD M SD M SD M SD F p
Graphic 60.00 17.96 47.22 11.02 57.57 14.35 62.03 15.65
1.03 0.3121
Visual 46.22 13.77 41.48 8.01 49.69 11.01 48.88 15.63
0.74 0.3926
Verbal 40.66 13.82 39.25 7.77 40.45 9.55 35.55 11.05
1.15 0.2863
Total 48.96 11.99 42.65 6.43 49.24 9.17 48.82 12.28
1.53 0.2191
C.3. - Aggregated Scores for the Representation of Three Emotions (See Table5.1 & Appendix N)
F-Test results indicate that there are no significant differences across the groups (F=1.55, p=0.
2168) when examining the aggregated scores of the representation of three emotions. The
comparison across four groups reveals that all students, except the Spanish monolinguals in the
two-way bilingual program scored similarly. Bilingual students in the two-way bilingual program
scored highest (M=49.18%, SD=9.18), followed by the English monolingual students in the
monolingual program (M=48.96%, SD=11.99) and bilinguals students in the monolingual program
(M=48.82%, SD=12.28). Spanish monolingual students in the bilingual program scored the lowest
(M=42.56%, SD= 6.43).
C.4. - Four-Group Performance In The Representations Of Three Distinct Emotions (See Table 5.1
& Appendix N)
Examining students’ performances in the different emotions provide us with a detailed illustration
of how the connection between levels of cognitive complexity and emotion.
Happy. F-Test results indicate that there are no significant differences across the groups
(F=0.17, p=0.6810). Bilingual students in the monolingual program (M=49.44%, SD=6.97) scored
the highest when representing a happy occasion, followed by English monolingual students in the
monolingual program (M=47.16%, SD=10.67) and by bilingual students in bilingual program
(M=44.35%, SD=13.08). Spanish monolingual students in bilingual programs scored the lowest
(M=39.53%, SD=10.82).
Sad. F-Test results indicate that there are no significant differences across the groups (F=1.
12, p=0.2737). Bilingual students in the bilingual program (M=52.30%, SD= 12.66) scored the
highest when representing a sad occasion, followed by English monolingual students in the
monolingual program (M=48.11%, SD=16.79) and by bilingual students in the monolingual
program (M=47.96%, SD=18.61). Spanish monolingual students in the bilingual program scored
the lowest (M=44.07%, SD=5.78).
Angry. F-Test results indicate that there are no significant differences across the groups
(F=1.57, p=0.2137). English monolingual students in the monolingual program (M=51.61%,
SD=16.70) scored the highest when representing an angry occasion, followed by bilingual students
in the two-way bilingual program (M=51.02%, SD=12.68) and by bilingual students in the
monolingual program (M=49.07%, SD=15.85). Spanish monolingual students in the two-way
bilingual program scored the lowest (M=44.07%, SD=11.81).
Table 5.1-
Average Ratings Across Four Groups & Across the Representations of Three Emotions
English Monolingual Speakers in Monolingual Program
(MM) Spanish Monolingual Speakers in Bilingual Program
(MB) Bilingual Speakers in Bilingual Program
(BB) Bilingual Speakers in Monolingual Program
(BM)
ANOVA
Variable M SD M SD M SD M SD F p
Happy 47.16 10.67 39.53 10.82 44.35 13.08 49.44 6.97
0.17 0.6810
Sad 48.11 16.79 44.07 5.78 52.30 12.66 47.96 18.61
1.21 0.2737
Angry 51.61 16.70 44.07 11.81 51.02 12.68 49.07 15.85
1.57 0.2137
Total 48.96 11.99 42.56 6.43 49.18 9.18 48.82 12.28
1.55 0.2168
D. – Comparison of Students’ Performances Based on their Enrollment In Two Distinct Educational
Programs (See Table 6 & Appendix O)
This comparison includes two groups:
Bilingual and Spanish monolingual students enrolled in the two-way bilingual immersion
program regardless of their home language are Spanish or English home-speakers who are being
schooled in two languages (n=53). The monolingual students in this program are recent arrivals
from Spanish speaking countries, which have access to entry in this program at any given grade
level; and
English monolingual students in the monolingual program with English-only as their home
language and bilingual home-speakers enrolled in the same English-only program (n=39).
D.1. - Students’ Performances In Three Modes Of Representation
(See Tables 6 & Appendix O)
The F-Test results indicate that there is no statistically significant difference in the means (F=1.18,
p=0.2809) when examining students the overall performance in the three modes of representation
(graphic, visual and verbal). Both students in the monolingual program (M=48.93%, SD=11.89)
and students in the two-way bilingual program (M=48.12%, SD=9.07) scored very similar to each
other. The lower scores of Spanish monolingual students in the two-way bilingual program that
we saw earlier, who are recent arrivals to the United States, contribute to this difference. In
further analysis, we are able to see that only the scores in students’ graphic representations affect
the difference. Results are presented in table 6 (Appendix P)
Graphic. F-test results determine that there is a significant statistical difference between the scores
of students in both programs in the complexity of graphic representations (F=4.44, p=3.79).
Students in the monolingual program (M=60.47%, SD=17.28) scored higher than students in the
two-way bilingual program (M=55.81%, SD=14.30).
Visual. F-test results determine that there is no statistically significant difference between the
scores of both programs in the complexity of graphic representations (F=0.37, p=0.5442).
Students in the two-way bilingual program (M=48.30%, SD=10.95) scored higher than students
in the monolingual program (M=46.83%, SD= 14.05).
Verbal. F-test results determine that there is no statistically significant difference between the
scores of both programs in the complexity of graphic representations (F=0.35, p=0.5541).
Students in the two-way bilingual program (M=40.25%, SD= 9.21) scored only slightly higher than
students in the monolingual program (M=39.48%, SD=13.27). There are no significant statistical
differences.
Table 6 –
Comparison of Two Educational Programs on Three Modes of Representation
Monolingual Program
N= 39 Two-Way Bilingual Program
N= 53 ANOVA
Variables M SD M SD F p
Graphic 60.47 17.28 55.81 14.30 4.44 0.0379
Visual 46.83 14.05 48.30 10.95 0.37 0.5442
Verbal 39.48 13.27 40.25 9.21 0.35 0.5541
Total 48.93 11.89 48.12 9.07 1.18 0.2809
D.2. - Students’ Performances In The Representations Of Three Emotions (See Table 6.1 &
Appendix O)
The F-Test results indicate that there is no statistically significant difference in the means (F=1.24,
p=0.2691) when examining students’ overall performance in students’ drawings associated with
three different emotions. Students in the monolingual program (M=48.93%, SD=11.89) and
students in the two-way bilingual program (M=48.06%, SD=9.07) scored virtually the same. In
further analysis we are able to see that differences in the representation of emotions are minimal.
Results are presented in table 6.1 (Appendix Q).
Happy. The F-Test results indicate that there is a statistically significant difference in the
means (F=4.26, p= 4.19). Students in the monolingual program scored higher (M=47.69%, SD=9.
90) than students in the bilingual program (M=43.53%, SD=12.76).
Sad. The F-Test results indicate that there is no statistically significant difference in the
means (F=0.00, p=0.9684). Students in the bilingual program (M=50.84%, SD=12.11) scored
higher than students in the monolingual program (M=48.07%, SD=16.97).
Angry. The F-Test results indicate that there is no statistically significant difference in the
means (F=0.54, p=0.4626). Students in the monolingual program scored higher (M=51.02%,
SD=16.34) than students in the bilingual program (M=49.84%, SD=12.70).
Table 6.1-
Comparison of Students in Monolingual and Bilingual Programs by the Representation of Emotions
Monolingual Program
N= 39 Two-Way Bilingual Program
N= 53 ANOVA
Variables M SD M SD F p
Happy 47.69 9.90 43.53 12.76 4.26 0.0419
Sad 48.07 16.97 50.84 12.11 0.00 0.9684
Angry 51.02 16.34 49.84 12.70 0.54 0.4626
Total 48.93 11.89 48.06 9.70 1.24 0.2691
Interpretation
The graphic, visual and verbal representations of children’s relationships of self and significant
others used for this study provided evidence of their cognitive levels. The significant differences in
the performance of all students across the three modes of representation point to the critical role
that visual assessments play in providing essential and complementary information about children’s
cognitive functioning. Children’s representations of Self in drawing captured the three categories
proposed by James (1983): the constituents of self, the feelings and emotions and the actions.
Symbols used by students to portray interactions in their drawings provided concrete quantifiable
evidence for analysis and therefore measure of cognition.
Results indicate that children’s thinking patterns evolve in stages as Piaget (1936) uncovered or in
levels as Fischer (1980) delineated. Children’s competence in mental operations, which are
reflected in actions and symbols, provide indicators for identifying their level of performance.
Results also showed evidence of the influence of socio-emotional development in children’s mental
capacity as suggested by Gunnar and Sroufe (1991). Drawings provided representational (visual
and verbal) evidence of the various levels suggested by Fischer and Mascolo (1998) as students
included persons on their own, a person in a relationship with another, and a person as it functions
in society at large (Appendix R: 1-16).
1). How do the cognitive structures of monolingual and bilingual speakers compare when
performing in three distinct modes of representation: graphic composition, visual content, and
verbal narrative?
Results indicate that there is no statistical difference between the cognitive structures of
monolingual and bilingual speakers as measured by the levels of complexity achieved in three
modes of representation. This finding is consistent with Piaget’s claim that linguistic experiences
have little or no impact on a child’s psychological development. This finding is also in line with
Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001) who found that bilingual and
monolingual subjects develop at similar rates in biologically controlled mechanisms related to
linguistic acquisition.
I feel secure in saying that the cognitive development of monolingual and bilingual youngsters
evolves, at least, at similar rates. Even though, I noticed a hint that bilingual speakers tended to
score slightly higher than their monolingual counter parts in most of the measures. A subsequent
study, with larger number of participants, would be necessary to confirm this trend.
In general, all students scored higher levels of complexity in graphic composition than in visual
content, though scores in both modalities were highly correlated. This slight difference may be due
to the fact that the graphic scale is based on elements of composition (such as horizon lines and
perspective) and it was a four-point scale (though the analysis counted for that difference) and the
visual scale was based on the nature of relationships represented in the content of the drawings.
The strong correlation between graphic and visual scores indicates that students perform similarly
when being assessed in the visual domain and points to the validity of the Varon scale.
The five-point visual and verbal scales used in this study are closely inter-related. Scores for both
scales are derived from content analysis, which assesses portrayed relationships based on criteria
designed by Fischer (1980) in his skill theory. Both visual and verbal scores show weak correlation,
which indicates that students’ performances in different modes/domains fluctuate.
My anticipation is that the Varon visual scale used in this study will improve the “traditional” tool
used to measure artistic development (i.e. Lowenfeld’s scale based on graphic composition derived
from Piaget stages in the 50’s) in the same way that Fischer revised Piaget’s measures of cognitive
development with his proposed levels. The Varon scale is a timely response to Feldman’s
suggestion that stages of artistic development be reformulated based on current theories of
cognitive development.
When examining the differences across four groups, the difference between the English
monolinguals in the monolingual program and the bilinguals in the bilingual program is 0.28 points
in favor of the bilinguals. The difference between the English monolinguals and the bilinguals in the
monolingual program is 0.14 points. Spanish monolingual students in the bilingual program, on the
other hand, score consistently lower. The gap between bilinguals and Spanish monolinguals in the
bilingual program is 6.17 points when looking at all three representational modes. This group of
Spanish monolinguals also scored much lower in their Raven Progressive Matrices. The gap
between them and the English monolinguals in monolingual program is 3.65. The gap between
them and the bilinguals in the bilingual program is 3.70.
This gap between the Spanish monolingual speakers in the bilingual program and the rest of the
groups is most likely the effect of their recent immigration. These students have not had the same
enriched environmental and educational exposure as the other students. The drastically changed
environment and temporary unstable living conditions may contribute to a psychological block of
expression. In time however, the bilingual educational program will prove beneficial to them, as it
has been for other immigrants who become proficient bilinguals in the two-way bilingual immersion
program.
Students in both educational programs scored virtually the same. When examining all three
representational modes, the difference is 0.81 in favor of the monolingual program, especially in the
graphic composition with a 4.66-point difference. This gap is evidently produced by the lower
performance of the Spanish monolinguals students in the two-way bilingual program who happen
to be new immigrants with lower Raven scores.
2) How do the cognitive structures of monolingual and bilingual speakers compare when examining
their representations of three distinct emotions (happy, sad and angry)?
This dissertation found that there are no statistical significant differences across monolingual and
bilingual speakers when representing any of the targeted emotions. This is valid also when
comparing across the four groups. Nevertheless, there is evidence of a statistically significant
difference when representing happy (p=0.0419) when comparing students in the monolingual vs.
bilingual program.
Bilingual speakers score higher than monolingual students when representing a sad or an angry
event. The difference is higher when representing sad (4.36 points) than when representing angry
(0.82 points).
When representing a happy event in all three modalities, monolingual and bilingual speakers scored
virtually the same (0.18 points difference in favor of monolinguals). Angry events produce the
highest level of complexity in verbal narratives for both monolingual and bilingual speakers.
In examining the representation of emotions across four groups, “sad” produces the highest score
by the bilingual speakers in the two-way bilingual program. Next is “angry” which produces similar
scores by English monolinguals in the monolingual program and bilinguals in the two-way bilingual
program with a difference of only 0.59. Spanish monolingual students in the bilingual program
score consistently lower when representing all three emotions, with “happy” scoring the lowest.
When looking at the representation of emotions across two educational programs, the difference is
0.87 also in favor of the monolingual program. This difference is consistent and being affected by
the lower performance of Spanish monolingual students in the bilingual program.
3). How do monolingual and bilingual speakers compare in degrees of cognitive complexity as
manifested in the graphic composition of their drawings?
There is no evident statistical significant difference in the artistic development of monolingual and
bilingual speakers as measured by the analysis of their graphic compositions. Bilingual speakers
scored slightly higher (1.28 points higher) than monolingual speakers. There are also no statistical
significant differences found across the graphic representations of the four groups. Nevertheless,
there is evidence of a statistically significant difference (p=0.0379) when examining students’
graphic composition and comparing students in the monolingual vs. bilingual programs. This is due
to the lower performance of the Spanish monolingual students in the bilingual program, who, as
mentioned before, come to this country with a disadvantage. This finding supports evidence of
Feldman’s theory of universal vs. non-universal development where he states that there is no
guarantee that a child will move from stage to stage at the same rate as others if he or she has not
had similar environmental and educational enrichment. Since non-universal development is highly
dependent on environment and individual resources, immigrant children tend to lag behind in
development for as long as five or six years. This is line with the time that Cummins (1981), Collier
(1987) and Hakuta (1987) prescribe necessary for children to acquire full academic proficiency in
two languages.
4). How do they compare in the degrees of cognitive complexity manifested in the visual
representation of interaction in their drawings?
There is no evident statistical significant difference between the cognitive structures of monolingual
and bilingual speakers when examining any of the comparisons of students’ visual representations.
Even if the difference is not statistically significant, when the group is broken down into four
however, the monolingual speakers in the bilingual program scored the lowest with an 8.21-point
difference when compared with the bilingual speakers in the same program, and with a 4.74-point
difference with the monolingual speakers in the monolingual program.
Additionally, the fact that bilingual speakers scored (4.43 points) higher than the monolingual
speakers using visual assessment and (1.96 lower) than when using verbal measures of assessing
(see below) indicates how crucial it is to diversify the domains in which we assess. Visual measures
provide valuable and complimentary information.
When examining the scores in the visual representations across four groups, bilingual students in
the bilingual program and bilingual students in the monolingual program score similarly with 0.81-
point difference in favor of bilinguals in the bilingual program. The widest gap is between the
bilinguals and Spanish monolinguals in the two-way bilingual program again, with an 8.21-point
difference in favor of bilinguals in the bilingual program.
5). How do they compare in the degrees of cognitive complexity manifested through the verbal
descriptions of depicted interaction as measured by skills theory?
Findings indicate that there is no significant statistical difference between monolingual and bilingual
speakers’ cognitive development based on verbal measures. Monolingual speakers scored only
0.72 points higher than bilingual speakers. .
In examining data derived from students’ verbal narratives across four groups, I noticed a 5.11-
point difference between the English monolingual and the bilingual students in the monolingual
program. This finding supports Lambert’s (1987) finding that subtracting bilingual speakers of
their native language in school will negatively affects their cognitive development. Conversely, the
evidence points to the fact that schooling bilingual speakers in two languages will promote their
verbal development in the least.
When comparing both educational programs in students’ visual and verbal scores, students in the
bilingual program have the advantage. In the visual representations students in the bilingual
program are ahead by 1.47 points, and in their verbal narratives they are ahead by 1.74 points.
Conclusions
Since Fischer’s (1980) claim that development is quantifiable even if multifaceted and divergent
according to domain and individual differences, skill theory provided a suitable developmental
system (scale used to measure verbal narratives) for studying monolingual and bilingual children’s
cognitive profiles as measured by the narratives that described their drawings. The visual scale
used in this study was based on a system of role intersections and relationships derived from
Fischer (1980) levels of cognitive development. Children’s levels of performance started with
simple concrete structures and moved towards greater degrees of complexity that included
symbols and abstractions.
The scale used to measure graphic composition in this study was based on Lowenfeld’s stages of
artistic development. Due to the differences between the artistic (graphic) and cognitive conditions
(visual content and verbal narratives), I predicted greater advantage in visual and verbal
representations, rather than in graphic composition. However, results suggest that there are
significant differences in students’ performances in the three different modes, with evidence of
strong correlation between the graphic and visual scores. This leads me to conclude that the
Varon scale offers a reliable alternative, and possibly an improvement, for measuring cognitive
development through drawing.
My objective in this dissertation was to apply quantifiable measures used in the field of cognitive
psychology and to contribute to the field of artistic development by proposing an alternative scale
to Lowenfeld’s stages, which are used to measure graphic composition. The bilingual vs.
monolingual context provided me with an opportunity to further look into the impact of linguistic
development on children’s psychological and educational experiences.
Initially it was surprising to note the slight disparity between the representations of cognition in
psychology and in bilingual literature. In bilingual literature, cognition is closely tied to academic
performance, whereas in psychology, cognition is tied to a child’s biological maturation of ability to
perform a thinking task.
Bilingualism is developmental in that it takes 5 to 7 years for students (who do not grow up in
bilingual homes) to become bilingual. In this case, bilingualism fits into the “non-universal”
category proposed by Feldman (1985), and its development is dependant on environment and
educational experiences. When researchers on bilingualism such as Collier (1987, 1995 a- b),
Cummins (1981), Hakuta (1986, 1987,1990, 2001) Lambert, 1972), Nicoladis, E. Taylor, D.,
Lambert, W.E. & Cazabon, M. (1998); Thomas & Collier (2000-2002) write about the cognitive
advantages of bilingualism, they refer to context-specific school-related tasks not necessarily
associated with biological maturation. On the other hand, psychologists such as Piaget, in his
stages of cognitive development, refer to the universal biological maturation, which determines
children’s readiness to be able to perform a mental task.
Fischer combines the various points of view, integrating biological maturation, environmental
influences, and context specific experiences in a task to determine his levels of cognitive
performance. These are domain specific and conditioned by biological structures. However,
interactive experiences and probing from an expert will allow for optimal performance in a specific
task. “Skill theory clearly states that specific skills develop as a result of experience in particular
tasks (although there is a biological basis for the optimal level that a child can reach)”. (Fischer,
2004)
The findings of this study confirm that cognitive development precedes linguistic experiences and
that it is therefore universal, as Piaget had claimed. There is also evidence that the environmental
(cultural) influences have significant impact on cognitive development. Spanish monolingual
students who were recent immigrants to the United States often performed at lower levels of
complexity than other groups participating in the study.
Whereas bilingualism may not propel psychological or biological advantages, it also does not in any
way cause cognitive deficit, such as mental confusion. This finding is supported by studies
conducted by Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001).
Results suggest that there are no evident statistically significant advantages to bilingualism as
measured by all three representational modes. There is only a hint at actual cognitive
development advantages because bilingual speakers scored slightly higher than their monolingual
counterparts in most of the instances. In order to further explore this trend, a follow up study
with an increased number of participants would be necessary. In any case, this study revealed that
bilingualism does not in any way hinder individuals’ verbal or cognitive development.
In this dissertation, I demonstrated the developmental appropriateness of the Varon visual scales
in that they produced scores that corresponded with Fischer’s levels for the verbal narratives, and
where there was a significant statistical difference in the Raven Progressive Matrices, there was also
a statistically significant difference in the results produced by the visual scale.
In considering children’s differences in students’ skills performance associated with drawing, I
explored practical applications of drawing in measuring cognition. Drawings have not been used to
examine the differences between monolingual and bilingual speakers, except in my qualifying
paper. In that paper, I suggest that in the process of learning a second language, while not always
capable of vocalizing thoughts, students become more observant of visual elements such as
setting, interaction, and personal characteristics of people and environments (Varon, 2000). This
may have direct impact in the perceptual system (Eisner, 1967; Mehler, Pallier & Christophe, 1998;
Varon, 2000; Whorf, 1956).
Additionally, I explored how traditional measures of artistic development, such as the one created
by Lowenfeld, relate to measures of cognitive development. The fact that children’s scores in visual
content correlate with scores in graphic composition demonstrates that visual content as well as
graphic compositions successfully capture cognitive levels of performance in the same quantitative
way as do verbal narratives.
The implications of this study could impact the assessment processes used in schools today. This
study may be relevant to teachers who want to better understand their students as well as the
benefits and constraints of being monolingual or bilingual. Drawing is a natural activity for children
providing a comfortable assessment tool. The task and framework for analysis used in this study is
one that can be easily implemented in the classroom on a regular basis; it would be an alternative
way of evaluating children who may not perform as well in verbal or written assessments.
Due to the fact that I did not collect academic scores of students participating in this study, I am
not able to differentiate between academic and psychological performances of bilingual and
monolingual speakers as other researchers have done (Cazabon & Lambert, 1994; Cazabon,
Nicoladis & Lambert, 1998; Thomas &Collier, 1997, 2002). However, I am able to conclude that it
is crucial to diversify the domains in which we assess children. Using visual assessment in addition to
verbal measures provide valuable and complimentary information about children’s cognitive
structures.
Suggestions for future studies include furthering the application of the visual representational scale
to conduct a longitudinal developmental study that would include the drawings and academic
assessments of monolingual and bilingual speakers in elementary and middle schools.
IMPRINTS OF SELF
Manifestations Of Bilingual And Monolingual Children’s
Cognitive Structures In Drawings Of Self And Significant Others
by Corinne Varon, Ed. D. Harvard University
It is my conviction that each drawing is the reflection of the personality of its maker; that it
expresses affective aspects of the personality as well as cognition… that the approach of the
examiner must of necessity be intuitive as well as analytical. (Di Leo, 1983, p. 60).
For my dissertation, I conducted a comparative study of drawings to describe how the nature of
self in interaction with others differs when produced by distinct groups of English monolingual,
Spanish monolingual, and bilingual speakers enrolled in two different elementary school programs.
Three drawing tasks were administered to all 5th and 6th grade students enrolled in one public
school in Massachusetts.
This exploratory study examines the cognitive structures underlying the construction of drawings
that capture representations of self and others. I examined the complexity of children’s cognitive
structures as represented in their graphic, visual, and verbal forms. This study also tested the
practical applications of a visual scale derived from cognitive measures based on behavioral
observations and verbal narratives. Children’s autobiographical visual and verbal recollections and
events associated with happy, sad, and angry supplied the data.
Davis (Lawrence-Lightfoot & Davis, 1997) affirms how art is the means to understand thought in
The Art and Science of Portraiture,
Through internal symbols or representations that are vehicles of thought, the producer of art
constructs a world view (p. 29).
Art allows children to reflect about their life, giving it a meaning while they communicate it to
others. Visual arts manifest ideas and feelings that may not be revealed through words or actions.
“Art puts into work, energies and contradictions that we would not otherwise be able to tap into"
(Lima, 1997, lecture).
To systematically compare performances, 92 subjects were selected. I compared up to 4 distinct
groups of students ages 10-13, grouped by their linguistic classifications and/or program
enrollment. Groups were also matched by their level of intelligence as measured by the Raven
Progressive Matrices scores (Raven, 1986).
Drawings were scored according to established ranked criteria in three categories graphic
composition, visual representation, and verbal narrative. The scales for graphic composition and
verbal narrative have been widely used to measure artistic and cognitive development. For visual
representation, I derived a scale from Fischer’s skill theory (1980) and adapted criteria for
identifying various levels of cognitive development through drawings (Table 2, Category 3, on page
62).
Regression analyses, using dummy variable coding, indicated that there were no major statistical
significant differences between the students grouped as English monolingual, Spanish monolingual
or bilingual speakers or as attending monolingual English or bilingual Spanish/English educational
programs. Correlations between scales and expressive processes indicate the reliability of the
measures across the verbal and visual domains in the performances of each group.
It is clear that in examining differences between monolingual and bilingual speakers, researchers
have applied numerous measures of intelligence, creativity, academic achievement, and socio-
metrics. The use of drawing has been relatively rare, if not absent.
My argument is that visual representations hold evidence of children’s levels of cognitive
complexity, particularly when analyzed for their content rather than their graphic skills in
representing perspective. The effects of “bilingualism” and the impact of dual linguistic
development on children’s cognition are imprinted in their drawings. The rational for selecting this
particular age group was that by middle childhood and pre-adolescence, there is evidence of the
manifestation of children’s skills in different domains.
During middle childhood, cognitive acquisitions allow the child to differentiate his/her abilities
across domains (Harter, 1999, p. 9).
In addition, by ages 11-13, most of the bilingual speakers in the study would have been in the
bilingual program for over five years. According to Cummins (1981) it takes over five years for a
monolingual child to become academically fluent in two languages.
By examining the correlation between established frameworks frequently used to map cognitive
development, and the proposed criteria used to identify imprinted cognitive structures as
manifested in children’s drawings, I hope to create a useful scale for analyzing the content of
children’s drawings and contribute a contemporary alternative to traditional research in artistic
development (Eisner, 1967; Eng, 1931; Goodnow, 1978; Golomb, 2002, 1972; Hildreth, 1941;
Piaget, 1983; Lowenfeld, 1939; Luquet, 1927; 1991 Thomas & Silk, 1980).
By examining the emergent patterns that diverse levels of complexity represented in different
mediums and for three distinct emotions, I am eager to reveal information to further our
understanding of children’s expressive processes and confirm the positive results found in two-way
bilingual immersion education in the United States.
As a bilingual educator with a strong background in visual arts, I have consistently considered
drawings to be manifestations of children's cognition. Using art, I have worked diligently to
developed pedagogies that increased children’s mental flexibility and creativity. Children's
contextual cues and structures of representation often reveal their understanding of self, the
nature of their thoughts, and their levels of cognitive development.
In fact, Koppits (1984) said, “During the elementary school years, boys and girls can express their
thoughts and feelings often better in visual images than in words” (p. 2). Selfe (1983), when
reviewing Susanne Langer’s (1953) writings, establishes a direct link between drawing and mental
experiences through the formulation of symbols.
Langer sees drawing as one of the expressions of man’s unique ability to reconstruct reality for
himself through conceptualization or, in her terms symbolization… She demonstrates that a
drawing is a formulation in symbols of the artist’s concepts and mental experiences (p. 23).
Arnheim (1974) also points to children’s ability to communicate in drawings in spite of the technical
challenges of reproducing what is perceived.
Children are technically unable to reproduce what they perceive… Much of the time, however, the
lines are accurate enough to indicate what the drawing is supposed to be like, particularly to the
observer who compares many drawings of the same kind (p. 163).
Literature Review
To further my understanding of methods used to measure children’s cognitive and artistic
development, and to compare monolingual and bilingual speakers’ relative performances in the
various domains, I examined three bodies of literature: 1) models of cognitive development, 2)
artistic development of pre-adolescents, and 3) effects of bilingualism.
Models Of Cognitive Development
Most psychosocial approaches to studying self reconstruct and examine its development through
subjects’ self-reports, which is considered a cognitive construct.
The self is, first and foremost, a cognitive construction... As a result, the self will develop over time;
that is, as cognitive processes undergo normative developmental change, so will self-concepts,
including their very structure and organization (Harter, 1999, p. 8).
In the twentieth century, the progression from theoretical to practical was induced by the
development of tools used to measure and define the personal self. In fact, Sandra Pipp (1990)
describes what has constituted development.
The type of understanding that can be brought to bear on self and other changes with
development, because changes in the structure of the thought influences how people process
information (Case, 1985; Fischer and Pipp, 1984) and so should also influence how people are able
to process information about the self. These cognitive structures serve as filters or upper limits on
processing capabilities (Fischer and Pipp, 1984, p. 253).
James, William (1842-1910)
James (1981) and Jung (1957) presented the theoretical foundation for the study of the self upon
which theorists covered in this literature-review have built. James introduced the idea that the self
is an internal operative entity composed of two parts (Me and I) with various constituents
involving a hierarchical structure (Damon, 1982; Harter, 1998; Fischer & Mascolo, 1998).
James (1983) defined the private, personal (Me), and public (I) selves in pluralistic terms and
proposed three categories for their analysis. These are as follows:
1. Constituents of the (Me) self (which individuals have arranged in a hierarchical scale according to
worth) represent the nuclear part of the Self:
a. Material Self: This self includes our body, possessions, clothing, family, home, etc.
b. Social Self: The multiple social selves are shaped by the roles we play. This is who we are,
how we behave in relation to others and how other see us.
c. Spiritual Self: This is the subjective part of us.
d. The Pure Ego: This involves our soul and the vehicle for the self-measuring process
2. Feelings and Emotions that the four constituents stimulate.
3. Actions encouraged by the four constituents (James, 1983, pp. 280-300).
James (1983) considered the (I) personal self in terms of individual functions and self-knowledge
as a continuously evolving entity involved in the process of awareness incorporating all
experiences, including the spiritual. The (Me) public self was based on societal recognition. He said,
“Man has as many social selves as there are individuals who recognize him and carry an image of
him in their mind” (pp. 282-282). According to Fischer and Mascolo (1998), the multiple selves
have “capacity to represent and regulate the self along multiple pathways, even within the same
individual” (p. 332).
In his Principles of Psychology, James (1983) proposed that “successive perceptions constitute the
mind” (Hunt, 1993, p. 333), and that the self is not a system of fixed relationships but rather an
ever-changing stream of consciousness (Hunt, 1993, p. 159). He also believed in the ability of the
self to “distinguish between the immediate and actual [objective], and the remote and potential
[subjective]” (James, 1983, p. 300).
As his principal method of investigation, James used introspective observation as a means to
revealing the states of consciousness. He explained,
Introspective Observation is what we have to rely on first and foremost and always. The word
introspection need hardly be defined - it means of course, the looking into our own minds and
reporting what we there discover. Everyone agrees that we there discover states of consciousness
(James, 1983, p. 185).
However, revealing the states of consciousness is challenging. According to Damon and Hart
(1982), “it is difficult to observe or characterize a phenomenon that is totally subjective and that,
therefore, may change unpredictably from moment to moment” (p. 845). Their claim was that
James was more of a philosopher than a psychologist, and did not provide a framework for a
developmental approach.
Although James admitted to some individual variation in how the constituents were formulated, he
did not recognize the possibility that their hierarchical interrelations might vary, both across
individuals and within one individual over time (Damon, 1982, p. 844).
According to Harter (1998), the symbolic interactionists Cooley, Mead and Baldwin contributed to
James’ theory by claiming that the self originated in the symbolic imitation of others and that it is
defined by how we internalize what others thought. They proposed that children develop a sense
of self through interaction when they internalize “the actions, thoughts and attitudes of others” (p.
555). Selfe (1983), when discussing Gibson (1979), shares this belief, “The pivot of Gibson’s
theory is that perception results from an interaction between the viewer and his environment”
(Selfe, 1983, p. 25).
Invariably, self- representations portray all aspects of Self (Me and I), involving them in
introspection, which is subjective. For this dissertation, subjects were asked to draw upon
introspections and represent the subjective. Before starting the research process I had to answer
the question of how to measure “cognition” based on the “content” which in drawing has
consistently been thought of as subjective.
My thesis is built upon the assumption that the choice of symbols used to communicate a
subjective message translates it into the objective realm, providing us with concrete agents or
elements, which can be used as data and can be analyzed quantitatively.
Jung, Carl (1875-1961)
Jung introduced the idea of basic character types, suggesting that the self passes through three
basic developmental phases: early, middle, and late life. He suggested that experiences are stored
subconsciously and that art productions reveal the whole self through symbols. Jung recognized
the importance of symbols. He specifically studied visual symbols, both produced and perceived by
his patients, and their application in the psychology of the self.
The self, I thought, was like the monad, which I am, and which is my world. The mandala
represents this monad and corresponds to the microcosmic nature of the psyche… The self is our
life’s goal, for it is the completest expression of that fateful combination we call individuality… [It
is] the individuation process of integrating the content of the unconscious and achieving
awareness of the self (Jung in Jaffe, 1979, p. 78 & 107).
Jung recognized the importance of others in the formation of self, claiming that the self without
relationships could not become whole. He stated,
Individuation has two principal aspects: in the first place it is an internal and subjective process of
integration, and in the second it is an equally indispensable process of objective relationship.
Neither can exist without the other, although sometimes the one and sometimes the other
predominate (Jung in Jaffe, 1979, p. 125).
James and Jung both offered theoretical frameworks that included constituents of the self, which
influenced all subsequent research during this century. Yet, they did not provide us with a
developmental trajectory of how the self evolves.
Nonetheless, Jung encouraged researchers to experiment with unprecedented innovative ideas
and methods, cautioning scholars about the dangers of looking at the self with existing
frameworks, whether scientific, anthropological or psychological. In an introspective moment Jung
declared how he comes up with a paradigm shift,
I must lay aside all scientific knowledge of the average man and discard all theories in order to
adopt a completely new unprejudiced attitude. I can only approach the task of understanding
with a free and open mind, whereas knowledge of man, or insight into human character,
presupposes all sorts of knowledge about mankind in general (Jung, 1958, p. 10).
In relation to drawing, contemporary researchers such as Koppits (1984) affirm that in the content
of children’s drawings, we can find non-verbal cues to the messages they send out about
themselves (p. 2) and that mental representations and events in children’s lives are direct
representations of children’s cognitive development (Gonzalez, 1999, p. 24).
Non verbal representations are perceptually similar to the events that they denote and they are
processed simultaneously. The dual-code theory also assumes that idiosyncratic and context-
dependent
connections between verbal and nonverbal representations exist and lead to a complex associative
network called referential connections (p. 24).
Vygotsky, Lev, (1896-1934)
Vygotsky “argued that culture [and language] actively constructs and largely determines both the
course and the degree of development (Feldman, 1999, p. 3).
The speech structures mastered by the child become the basic structure of his thinking…"there is a
fundamental correspondence between thought and speech in terms of one providing resource to
the other; language becoming essential in forming thought and determining personality features
(http:www.sk.com.br/sk-vygot.html Vygotsky and Language Acquisition, p. 1-3).
Based on the conception of what constitutes the understanding of “a collective” when organizing
puzzles with sets of blocks, Vygotsky postulated 3 phases of verbal thought, each with its steps
(Blunden, 1999, pp. 5-6; Vygotsky, 1994, pp.11-57):
1) Syncretism: objects are organized by their subjective connections
- Trial and error;
- Egotistical Selections made by trial and error based on the visual field; and
- Combinations of both trial and error and egocentric manifestations.
2) Complexes: Child is able to combine the subjective with facts but results are not always
consistent
- Associative Complex – manifested by what is related and familiar;
- Collection Complex – manifested by sets, collections with different attributes;
- Chain Complex – manifested by one object connected to another by common attributes but
that to the next by a different attribute;
- Diffuse Complex – manifested by variability of the attribute which results in an unstable
connection; and
- Pseudo Concept - manifested by the concrete, visible likeness and by the formation of
associative complex limited to a certain kind of perceptual link.
3) Concepts: A single attribute is identified as the basis of a collective
- Potential – in its simplest form
- Concept Proper – concepts emerge when attributes are integrated and abstractions are
formulated
Vygotsky suggests that phases are not clear-cut and that they overlap as children develop. This
idea continues to be scientifically demonstrated in neuropsychological research as development
gets mapped with images of the brain as it develops (Fischer, 1980).
Phases co-exist with each other in the course of development,
predominantly in different areas of activities and at different times and in different situations. At
the same time, each stage builds upon the gains of the earlier (Blunden, 1999, p. 7).
Luria, A. R. (1902-1977)
Luria, a contemporary of Vygotsky, examined the difference between "people with graphic
reflection of reality and those with predominantly abstract, verbal, and logical approach to reality.
He looked at the differences in how individuals abstract and generalize when asked to group and
classify objects. He claimed that perception of self is related to the perception of others.
The perception of oneself results from the clear perception of others and the process of self-
perception reshaped through social activity, which presupposes collaboration with others and an
analysis of the behavioral patterns (http://marxists.org/archive/luria/works/1976/problem.
htm pp. 14-15).
His study was based on the premise that "the essential features of mental processes depend on
the way they reflect reality; therefore a particular form of mental activity should correspond to a
particular level of reflection (http://marxists.org/archive/luria/works/1976/problem.htm, p. 14).
After World War II (WW2) 1939-1945
After World War II (WW2), psychological methodology incorporated to the view that the self was a
cognitive construction defined by external and internal factors. When examining and describing
how aspects of the self develop, contemporary theorists rely on behavioral observations, context,
and emotions, all influencing children’s interactions with others determining their thinking patterns
and cognitive development.
The influence of socio-emotional development changes in children's mental representational
capacity has received increasing attention in recent years (Gunnar & Sroufe, 1991, p. 126).
Socialization experiences in children's interactions with caregivers, peers, teachers, and in the wider
socio-cultural context will influence the particular content and valence of one's self-representations
(Harter, 1999, p. 9).
When describing how aspects of the self develop, contemporary theorists have relied on behavioral
observations emphasizing contextual factors and social influences that influence children’s cognitive
development.
To our understanding of the nature of social experience (or perception), the defining
characteristics of human sociality, distinctive features of the physical and social environment, and
the psychological nature of the family as the primary social structure within which children develop
(Wosniak, 1993, p. 77).
Piaget, Jean (1896-1980)
According to Piaget (1923, 1926), the development of knowledge, thinking skills, and mental
images (or schemes), as well as the construction of logical-mathematical thinking patterns is
comprised of stages.
Piaget’s (1960) stages, which come about by the child’s interaction with his biological and
contextual environments, have served as an enduring framework in developmental psychology and
in other domains.
Piaget argues that these stages in cognitive development are neither biologically programmed nor
given by the environment, but instead constructed or created by children as they actively attempt
to understand the world (Feldman, 1999, p. 59).
The stages are as follows (Flavell, Miller & Miller, 1993, p. 133):
Sensory-motor: (birth - 18 months)
Pre-operational – Symbolic (18 months - 7 years)
Concrete-operational – Relations and Logic of the Symbolic (7-11 years)
Formal Operational – Abstract (11 -15 years)
Piaget claimed that motivation for action or thought is intrinsic to the cognitive maturity or level of
development of the child (Flavell, 1984). After many years of clinical interviewing, noted
observations, demonstration, and experimental evidence of concrete actions that represent the
internalization of mechanical competencies, Piaget (1936) classified four successive stages of
children’s cognitive development. Competencies (Fischer, Bullock, Rotenberg, & Raya, 1993) are
determined by the interaction of skill levels and neural systems
Competence is an emergent characteristic of a person in context… The dynamics of changes in
competence are explained by the analysis of developmental levels of skills as well as a neural
network model (pp. 93-94).
According to this framework, children at ages 7-9 function at the pre-operational and concrete
operational stage. At the pre-operational level, children view the world subjectively. According to
Di Leo (1977), the conceptual way children draw the human figures tells us a great deal about this
intellectual capacity.
Human figure drawings express the child’s subjective view with total disregard of actual body
proportion. Instead there is an emphasis on what is regarded as important – intellectual not visual
realism (p. 36).
At the concrete operational stage, children are able to understand conceptual categories and are
able to understand how others experience the world. According to Di Leo, “concept now
dominates perception” (1977, p. 36). Gardner (1994) says that children at this stage are able to
function at a symbolic level,
[They are] capable of a series of mental operations that constitute a structured whole… The
operations are actions performed mentally; in our terms, the making system is functioning on the
symbolic level (p. 302).
Fischer and Mascolo (1998) established four distinct levels of operation for the self. These are: a)
the biogenetic level which includes the biological functions of the body; b) the organismic agentive
which is the level that includes the person on its own; c) the social-dyadic level which includes all
relationships with one other individual; and d) the cultural-linguistic level which includes the person
as it functions in society at large.
In addition, as children develop various capacities in separate domains (Gardner, 1982, 1983;
Fischer & Mascolo, 1998), some theorists have paid greater attention to the conceptual structures
underlying the identification and meaning attributed to experiences.
The construction of the experience requires the detection and attribution of meaning to
information from the environment. Detection and meaning attributions are dependent of the
development of conceptual structures and processes in which information from and about the
environment is actively gathered, assimilated to appropriate concepts, and thereby interpreted
(Wosniak, 1993, p. 79).
Feldman, David
Feldman (1998) emphasized the distinction between the universal and the non-universal in
development, stressing that both internal and external forces come into play when a child
develops. Feldman (1985) called universal development the inevitable progress that all humans
make as they grow, requiring no special effort, instruction or environment. On the other hand,
non-universal development is the transformation that is highly dependent on environment and/or
individual resources. He claimed that the “movement from stage to stage is not guaranteed in
domains other than universal ones” (1985, p. 84). Feldman (1994, 1999) critiques some theories
of artistic development, Lowenfeld’s in particular. He says that these theories mix universal,
cultural and discipline based factors and suggests that stages of artistic development need to be
reformulated based on current theories of cognitive development.
This dissertation reformulates the traditional measures used in figuring out the “old” stages of
artistic development and identifies “new” criteria based on established Fischer’s (1980) levels of
cognitive development.
Fischer, Kurt
Fischer’s framework has been applied to developmental studies of self and identity development
across cultures (Cheng, 1998; Pipp, Fischer & Jennings, 1987; Karcher, 1995; Kennedy, 1994).
Data on complexity of conceptions of self are collected and these reveal participants’ levels of
cognitive development. In these studies, participants assigned descriptors and numeric values to
their significant relationships, placing them on circular diagrams. Participants grouped descriptors
according to similarities and differences, thereby determining their level of development
corresponding to Fischer’s measure.
In cognitive psychology, Fischer (1980) challenged Piaget’s framework when he discovered that
disruptions in development happen more frequently than Piaget had proposed and that these
disruptions happen in an individual’s performance of a specific skill in relation to context.
Performances take into account the role of context – therefore are domain dependent.
In his dynamic systems theory, Fischer (1980) delineates 13 levels of cognitive development
(Appendix A), overlapping one another at times. He asserts that children develop their cognitive
abilities, thereby internal representations, in incremental patterns progressing from simple to
complex (Harter, 1998, p. 569).
Representational internal working models are maps of the interpersonal environment which
provide individuals with predictions to understand themselves and others... The representational
internal working model is accessed consciously
(Pipp, 1990, p. 259).
Fischer (1980) examines the organization of behaviors in a wide range of skills proposing a
“recurring cycle of four levels: reflexes, actions, representations, and abstractions. He states that
the self undergoes development in different domains, each at its own rate. The self is represented
by its capacities along multiple contexts.
Capacities to represent and regulate the self develop along multiple pathways, even within the
same individual. As such, the self is not a unified entity that emerges at any single point in
development. Rather, the individual consists of a series of weakly connected control systems that
represent and regulate different aspects of self (Fischer & Mascolo, 1998, p. 332).
Skills vary not only between people but also across contexts for a given person (Fischer, Bullock,
Rotenberg, & Raya, 1993, p. 96).
Fischer (1980) describes his levels of development built on predictable changes on the nature of
complexity of thought, which in turn induce action as children learn a skill (Pipp, Fischer & Jennings,
1987; Shaver & Lamborn, 1887; Fischer, 1980, 1989; Fischer & Watson, 1980; Fischer & Pipp,
1984; Fischer & Bidell, 1992, 1997; Fischer & Rose, 1993; 1994; Fischer & Mascolo, 1998).
The cognitive processes through which structures relevant to incoming stimulus information are
accessed, through which that information is assimilated to the cognitive system, are acts of
meaning attribution. The cognitive system is a device for the generation of meaning (Wozniak,
1993, p. 79).
With development, self and other become increasingly differentiated because higher order
representations of categories of action are formed for self and other (Pipp, 1990, p. 255).
According to Fischer (1980) a cognitive structure is “a structure for knowing, a procedure that the
child actively applies to things in order to understand them” (p. 482). He states: “it is in the very
nature of intelligences that each operates according to its own procedures and biological bases” (p.
68). Skill refers to an individual’s capacity to behave and think within a given context and
psychological domain. A skill refers to an action performed at a fixed time in a specific context
(Fischer & Mascolo, 1998, p. 339).
Skills are defined in terms of environment as well as organism… The concept of skill, which itself
connotes a transaction… is a unit of behavior composed of one or more sets… The characteristic
structure of each level is a type of skill (1980, pp. 478-483).
A skill refers to an individual’s capacity to control behavior and thinking within a given context and
psychological domain. A skill is always an action performed in a context (Fischer & Mascolo, 1998,
p. 339).
Competence refers only to the upper limit on the developmental level in observable behavioral
patterns (Fischer, 1980; Fischer & Bidell, 1992, 1997; Fischer & Pipp, 1984; Fischer & Rose, 1993,
1994). Fischer’s approach examines children’s cognitive structures and their development through
self-representations, describing the relationships between two opposite behavioral factors (such as
nice and mean) during pretend play (Fischer, Hand & Hencke, 1993; Fischer & Rose, 1993) or
person and context (Wozniak & Fischer, 1993, p. 9).
Knowledge [is]...a structural mental code embodying information about the environment.
Knowledge is both given in the human biological endowment and extracted from experience
Wosniak, 1993, p. 83).
Children represent themselves in parallel relationships prior to reciprocal relationships, portraying
either positive only or positive and negative tangible features (Fischer & Mascolo, 1998, pp. 360-
361). For example, at the first level of complexity a student describes his drawing with a single
representation as follows: “I am alone in my house. I have my birthday party and I am having
fun. I am dancing and confetti is falling off and balloons are flying all around”.
At the second level of complexity, the representation involved mapping of two several items
focusing on one aspect of each: A child describes himself, his grandmother and his mother in the
kitchen in his house: “my grandma told me that my dog died and my mom is looking at my
brothers to see what they are going to do. There is a path that leads to where they are. I am
thinking how he used to look. He used to live with me in the Dominican Republic with my grandpa.
Grandma is talking with my mom and my mom is looking at my brothers are thinking that she may
have to say, excuse me I have to go check on my brothers. My grandma is eating next to the
microwave and there is a fridge, the sink stuff and a window that you can see outside”.
At more complex levels of interaction, children have underlying intentions motivating their actions.
This involves more complex systems of operation as children develop behavioral expectations of
others (Fischer, 1990, pp. 5-6).
Interactions differ from other actions in the nature of the expectational control that the cognitive
system exerts as the interaction unfolds (Wosniak, 1993, p. 82).
At the third level of representation the child involves relating several aspects of each item in a
system. He describes herself and grandmother in Cambridge at the funeral home: “I am crying
because my grandma died. I am thinking it was my fault because we went places when she did not
have a coat on and it was cold. She was really sick and we still went out”. Finally, at the fourth
level of representation, the child coordinates several concrete inter-relationships (systems) in terms
of an abstract or intangible characteristic. For example, a child draws himself alone at his house
when he was locked out: “I am waiting for my brother to come because I forgot my keys and
sometimes it takes a really long time so it makes me really mad because he is supposed to come
home early. He takes my keys everyday when he is not supposed to because he lost his own keys.
He gets up earlier than me and I think I can’t find the key, but it is that he has it. I am thinking,
when is my brother going to be home?”
Exploring these approaches further, in this dissertation Fischer's framework has been examined as
it applies to drawings by allowing subjects to draw their relationships.
Application Of Fischer’s Framework To The Proposed Study
Experiences get translated into meaning and in order to represent those meanings to communicate
with others, humans create symbols. Symbols are individual in their nature, but collective in their
use.
In addition to experiencing the world, human beings also generate verbal and imaginal symbols to
stand for it. Semiotic activity, the mental activity of symbol generation, has both representational
and cultural functions. Representationally, symbols stand for something other than themselves.
Through symbol, we can bring to awareness objects and events not currently present to the
senses.
Culturally, symbols embody the historically derived system of social meanings held in common by
members of the broader society. Much of human ability to communicate interpersonally, depends
on the fact that we possess shared systems of symbols (Wosniak, 1993, p. 80).
Fischer's framework (1993) is well suited for my study because of the symbolic congruence that
exists between play and drawing, and its potential to contribute to our understanding of
development levels in children’s drawings. In both activities, children rely on the interaction of
perceptual and conceptual processes creating a symbolic representation that parallels reality
(Krampen, 1991). Piaget (1983) also correlated play, which is based on imitation of real life
situations, with the formation of symbols.
Symbolic play… always involves an element of imitation functioning as the significant, and early
intelligence utilizes the image in like manner, as a symbol or significant (p. 126).
Fischer’s established framework to map skills development can be used to measure the nature of
represented interactions and relationships in drawings, especially since developmental sequences in
drawing that have been proposed for pre-adolescence (Di Leo 1973; Golomb, 1972) are not
specific enough to describe the types of disruptions identified by Fischer and colleagues in cognitive
psychology and neurology. Based on Fischer’s developmental progression, children are gradually
able to distinguish social roles and their complexities. Children begin with the understanding of
relationships that are similar in nature, followed by understanding interactions that are opposite,
sequential, reciprocal, instrumental, and collective.
Skill theory provides a suitable developmental system for studying children’s cognitive profiles
because Fischer (1980) claims that development is quantifiable even if multifaceted and divergent
according to domain and individual differences. This study explores how the representation of
cognitive structures can be quantified and interpreted into levels of complexity deduced from
children’s visual and verbal symbols.
WATSON, Malcolm.
Watson (1990), in his analysis of role-play proposes parallels steps to Fischer’s levels. He describes
the rules of dynamic transformation of interaction in particular when children enter the age of
representations, when they are able to formulate their actions according to social motivation and
to coordinate various roles.
A social role is defined as the cluster of prescribed and expected behaviors that pertain to a
particular category of people. The cluster of behaviors for one role is determined by the
relationship of that role to complementary roles. Roles are complementary when they are normally
associated and functions of one cannot be described without reference to the other (Watson,
1990, p. 271).
Watson’s steps begin with the differentiation of agents rooted in concrete behaviors (comparing
similarities and differences between agents) and progress to the integration of the roles based on
social relations and eventually on networks of relations in preadolescence. Like Fischer, Watson
points out that "it is important that the tasks be related to the specific domain in which one is
testing skill development" (p. 274).
Watson’s model is useful in providing details that I considered in the analysis of drawings seeing
that his steps 2-6 correlate with Fischer's levels of representational systems:
Step 2: Self as Agent - Social Role (Corresponding to Fischer’s Rp1) – Foci of behavior is on
the relationship between two complementary roles;
Step 3: Passive-Other Agent - Shifting Social Role (Corresponding to Fischer’s Rp2) – Child
is able to understand that one person’s role can shift, but does not understand that someone
could occupy the two roles simultaneously;
Step 4: Active Other Agent - Role Intersection (Corresponding to Fischer’s Rp3) – Child is
able to integrate the roles of self and others by bearing in mind multiple and simultaneous role
relations; and
Step 5: Integration of Roles - Role Network (Corresponding to Fischer’s Ab1) – Child is able
to compound several role intersections and can think about combining them to form a network;
and
Step 6: Network Relations (Corresponding to Fischer’s Ab2) – Children are able to relate
one network with another.
To sum up, children first imitate the actions of others, portraying emotions as either positive only
or negative only prior to combinations of positive and negative (Fischer & Mascolo, 1998, pp. 360-
361). At the next level, children are able to oppose, remedy, and reciprocate the action of others,
as well as use others as instruments to perform self’s desired actions. As they are able to count on
others to perform what the self cannot accomplish independently, they get involved in role
intersections where self is part of a compound action and others are recipients. Based on this
system of role intersections, and network relations, self becomes aware of shifting social roles
(genetic, physical, and emotional), where self is seen as an agent of actions that others may or may
not complement or support. Gradually, children begin to grasp abstractions as they develop
deeper understanding of collective actions. Finally, through abstractions, they are able to express
social or personal motives through acceptable symbols, which generally occur during adolescence
(Cicchetti & Beeghly, pp. 267-279). At these more complex levels, children’s actions manifest
hypothetical ideas such as justice, conformity, intimacy, etc. (Fischer, 1990, pp. 5-6).
Fischer and Mascolo (1998) summarize the overall agreement on the theories of self when they say
that the self develops as the physical body interacts with family values, society, culture and
historical context.
Selves are not decontextualized, abstract, or disembodied entities. Rather, selves are constructed
in particular bodies within particular physical, social, cultural and historical contexts. The
production of affective and bodily experiences consists of an especially important source of the self’
s embodiment (p. 350).
The self develops influenced by both internal and external factors. The development of self can be
observed, measured and analyzed. As methodology evolves research has the capacity to be more
accurate and reliable. As long as children’s construction of their self-concept is measurable
(starting simple and moving to complex) and that the representations of self reflect levels of
complexity in the emerging patterns, it is feasible to compare two groups of children with distinct
cognitive experiences.
The design of the drawing tasks used as data for this dissertation included representations of the
self in its physical and emotional contexts associated with the lives of children. My hope is to
contribute to further our understanding of cognitive development through children’s visual
representations of self and others.
Artistic Development Of Pre-Adolescents
Nature Of Drawing
Across the board researchers state that children’s drawings represent the interpretations of
reality. Many believe that data drawn from drawings provide a source of understanding of children’
s psychological and emotional development.
Young children are not only indulging in self-expression in their drawings, but are representing real
relationships. Their drawing
can be checked against structural descriptions of the external world (Freeman, 1980, p. 37).
Drawings capture symbolically on paper some of the subjects’ thoughts and feelings. It makes a
portion of the inner-self visible… More is revealed by the content, which is largely determined by
the way the subject, consciously or unconsciously, perceives himself, and significant other people in
his life (Klepsch & Logie, 1982, p.6).
Drawing is, above all, the construction of meaningful patterns out of simple abstract elements
(Freeman, 1980, p. 55).
Freeman (1980) explores what constitutes a representation and identifies three components: a
definable object, definition of a relationship, and the study of action. “Drawings clearly represent
objects and people, and the relationship between them, and they got through a well defined
process of development” (p. 50-51).
The criterion for selecting the appropriate inclusive task in drawing was that drawing must capture
both process and product as well as reflect internal and external structures. This approach is well
supported by some researchers.
The very act of drawing involves many interacting processes, sensorimotor coordination,
perception of the external model when there is one, consulting inner image or schema or
representation in memory when no external model is present, production of a graphic form and its
correction, understanding of pictures a medium of representation and communication (Lange-
Kuttner & Thomas, 1995, p. 2).
Arnheim (1974) postulates that drawings are inclusive of both internal and external structures.
In thinking about why art work proceeds from the simpler to the more complex level, we realize
that internal as well as external factors must be considered. Internally, the organism matures, and
as it becomes capable of more differentiated functioning, it develops the urge to apply this
capacity. The development, however, is not conceivable without the external world, which offers
the whole variety of directional relations (p. 188).
Yet Arnheim, in Harris (1963), recognize the value of drawings as final products described as
voluntary statements about precise realities and particular characteristics.
The work of art is a statement about the nature of reality… From an infinite number of possible
configurations of forces, it picks and presents one… the final solution of the problem of how to
organize a reality pattern of given characteristics (p. 177).
Other researchers, such as Basset (1977) suggest that drawings are not the reflections of internal
representation. Basset designed a study to demonstrate that children’s concepts of self are
complete in spite of what their representations tell us because young children (4.9 years old) were
asked to draw a human figure, and then to construct a figure with given parts. Children most
often drew incomplete pictures, yet never missed using all the body parts to build a complete
figure. Basset explained that “the image is complete, but individual components are not clearly
differentiated” (p. 49), which led her to conclude that drawings do not represent children’s internal
representations. When scanning a picture, young children do it along a vertical axis, paying
attention first to top and then to bottom (pp. 58-59).
Drawing As Representation
Drawing, when intended, can reflect the artist’s reality of what is perceived and conceptually
understood. As verbal language develops from simple to complex, visual arts develop from
concrete to symbolic. When external stimuli interact with the internal mechanisms, they produce
concepts, relationships and systems. These get expressed through narratives, dialogue and
symbolic representations. Drawings are different from photographic images in that they reflect
concepts based on the perception of reality
According to Arnheim, children tend to develop concepts before they are able to find means of
representation.
Image making of any kind requires the use of representational concepts. Representational
concepts furnish the equivalent, in a particular medium, of the visual concepts one wishes to depict,
and they find their external manifestations in the work of the pencil (Arnheim, 1974, p. 169).
Freeman (1980 separates the process in two: projection and denotation systems. These are based
on spatial axes (linear perspectives or points of view), as he asserts that, “drawing gives a printout
of his conceptual store… that the child’s drawing is knowledge dominated… [and] that the child
can draw far more sensitively to visual models than appears, because he extracts the ‘structural
core’ of the model” (Freeman, 1980, p. 28).
Freeman (1977) suggests that observing process and sequence in which drawings are made is
more valuable than the final product because the final product tells us little about the child.
Convincing though post hoc analyses of finished products may be, they cannot substitute for
performance analysis of ongoing processes (p. 25).
Arnheim (1974) also emphasizes the importance of process as a component of research that
should not be neglected,
The order in which children produce the various parts of their drawing is most relevant to the
psychological meaning of the work and should not be neglected in research (p. 173).
I strongly disagree with Freeman’s statement that the finished product provides little information
about the child, yet I acknowledge the value of process. On the other hand Freeman’s
recommendation for designing research studies using art as a data collection tool is well
appreciated.
An image can be made to work if one devised a task, which gives the image a functional role to
play… One can devise studies in which people behave as though they organized their behavior
around some analog representation of a think in the outside world (Freeman, 1980, p. 42).
When looking at drawings as finished products it has been prevalent to look at the whole picture,
Karen McNiff (1982) declares that children communicate visually and that art represents a
synthesis of a child’s world beyond imitation.
It is not a process of imitating or copying the physical world, but rather of synthesizing life
experiences. Art then becomes a means through which the child can communicate about the
phenomena which are too complex to describe verbally, but which are being perceived and
integrated into the child’s organization of reality (p. 275).
The system used to analyze drawings in this dissertation starts with the examination of the finished
products in terms of the evident relationships of the parts that constitute that “synthesis” and
evaluate the nature of the relationships between those parts. Maturity is represented by the
degree of integrity with multiple layers (or systems) of relationships. When looking at students’
graphic compositions, the relationships are based on elements of perspective (point of view).
When examining visual content and verbal narratives, relationships are based on the nature of the
depicted interactions of self.
Drawings Of Self And Significant Others
The analysis of drawings of self and significant others was first introduced by Hulse (1951) [as
reviewed in Klepsch & Logie (1982, pp. 77-110)] with the “Draw-a Family” technique. Norman
Freeman (1980) defines self-image in drawings as a concept that integrates information from
diverse sources and defines choices about spatial location and actions.
The body image is supposed to be a sort of composite conception of one’s ‘self’ constructed from
information from different sources… Many people are convinced of its usefulness though,
since it is an appealing notion that drawings are connected with one’s conception of, and feelings
about, oneself. This could then provide a link between the representation of the self in highly
structured drawing and expression of the self… The body image [is] a construct used to explain
how we are enabled to take decisions about spatial location and the actions, which are necessary in
dealing with that information (p. 45).
Attention is paid to the degree of synthesis examining: closeness, positions, omissions of figures,
size relationships, etc., to determine degree of maturity, family dynamics, as well as cultural and
racial differences. Using the same approach, Hildreth (1941) studied a sample of 149 train
drawings, produced by children ages 2-11, for their graphic narration mostly achieved through
symbols and their story telling arrangements (p. 159). She found that “bright children tend to
show in their drawings more correct synthesis than others, and older ones more than younger
ones” (p. 151). However, she claims that development is not simply about synthesis, but rather it
is about “ progressing more and more toward creative, imaginative drawing, using habitual
schemas but constantly introducing original touches (p. 153). With the recognition that visual
representations need not to replicate all aspects of an object, omissions lose their influence as
measure of development. This type of analysis is frequently used today as diagnostic aids in
psychological testing (Kramer, 1971, McNiff, 1992) rather than to map cognitive development.
Theories that examine drawings as finished products originate with the idea that drawings are
projections and intimate representations of the self (Harris, 1963; Machover, 1949).
The figure drawn is related to the individual who is drawing with the same intimacy characterizing
that individual’s gait, his handwriting, or any other of his expressive movements (p. 43).
Klepsch and Logie (1982) wrote about the four projective uses of drawings; one of which reveals
how children perceive themselves in relation to others. They describe these drawings in the
following way,
Group drawings are useful if one wants to find out how a child perceives himself within a particular
group. When children draw themselves along with their family or friends or teacher or
schoolmates, they project into their drawings their view of themselves in relation to others in the
group (p. 12).
Stages Of Artistic Development
There is no controversy over the facts that the visual arts comprise of systems of meaning with its
own rules and developmental sequences; that drawing production is directly linked to mental
development (Arnheim, 1954; 1974; Di Leo, 1977; Eng, 1931, 1957; Golomb, 1972, 2000;
Goodenough, 1926; Harris, 1963; Kellogg, 1969, 1970; Klepesch & Logie, 1982; Koppitz, 1968;
Lowenfeld, 1939, 1957, 1970).
Artistic development was originally built on observation of changes that individual children
demonstrated as they invented ways to express their ideas in drawing.
Drawing development is largely accounted for by creating children’s discovery or invention of the
pictorial schemata and of procedures for creating these schemata on paper (Thomas, 1995, p.
107).
The documentation and interpretation of events and behaviors during children’s acquisition of
graphic, symbolic, and communicative competence has facilitated the demarcation of predictable
stages of artistic development (Eng, 1931, 1957; Davis, 1997, 1991; Di Leo, 1973; Duncum, 2000;
Fischer, Jennings, Lamborn, Pipp & Shaver, 1985; Gardner, 1980; Golomb, 1972; Kellogg, 1966;
Koppits, 1984; Korzenik, 1972; Lowenfeld, 1939; Matthews, 1999; Thomas & Silk, 1990; Varon,
2000; Winner, 1982,1986).
There is also widespread agreement that in the image lies in a conceptual understanding. This
understanding (therefore image) is simply expressed at the beginning levels and gets more
complex with experience and skill development. Consequently, a collection of drawings from
children at various ages and skill levels provides a developmental continuum, and each drawing a
snapshot of children’s thinking, preferences and skill level.
According to Harris (1963), Hevner (1935) and Walton (1936) were the first to demonstrate how
drawings capture and communicate children’s emotions developmentally and to compare this
communication to the structure of spoken language. “Lines and forms may express feelings as
stated in adjectives” (p. 39). Harris (1963) explains Walton’s finding that “children’s attribution of
feeling involves principally including the pleasant-unpleasant dimension. However, as they learn
more words and concepts, children make finer distinctions and discriminate more polar dimensions”
(p. 39).
Representation implies abstraction and simplification (Arnheim, 1954) as opposed to duplication.
Duplication demands precision, attention to detail and to multiple relationships, a developed special
framework, and the concept of measurement (Piaget & Inhelder, 1956). Progress in the ability to
copy geometric figures parallels to some extent the increasingly stringent demands of “likeness”
which the child sets for himself, advancing from the early simple forms of equivalence to more
complex and mutually regulated ones (Golomb, 1974, pp. 184-185).
Golomb (1972), agreeing with Arnheim(1974), affirms that children’s representations in drawing
evolve from simple to complex that visual language operates similarly to verbal language. She says,
Drawing is a uniquely human activity whose complex syntactic and semantic development can be
studied systematically… Arnheim calls attention to the general developmental course that
proceeds from simple to complex forms, a developmental principal of differentiation that
characterizes all cognitive processes (pp. 2 & 31).
Golomb (2002) explores the nature of representation and the role of internal models of mental
images. She establishes the difference between development that moves towards realism and the
one that moves towards the representation of diverse objects (p. 5).
Golomb (2002) also differentiates between representation and perception, expanding Piaget’s view
that representations are systems of differentiated symbols that stand for recognizable objects.
Her broader view explains representations as “inventions of pictorial forms that stand for the
intended objects” (p. 5). It is the mechanics, principles, and symbols children use to express their
ideas and emotions about themselves and others that are considered projections of what is in a
child’s mind. Development progresses from “topological relationships that ignore true shape and
proportion” to realistic depictions, intellectual realism and finally, photographic realism (p.11). Di
Leo (1983) implied the same view of development when he says, “These terms express, in
substance, a metamorphosis in thinking from egocentricity to an increasingly objective view of the
world” (p. 37).
Freeman and Cox (1985) propose a completely different view of development. They claim that
development moves from object centered to view centered.
Goodenough (1926) was the first scholar to demonstrate that through the study of drawings of
the human figure (HFD) one could measure children’s intelligence and examine the salient
characteristics of a child’s personality. Her measure (which was later revised to be the
Goodenough- Harris measure) provided many scholars with objective criteria with which to analyze
drawings, propelling numerous comparative studies of children’s drawings (Harris, 1963, p. 11).
As for the development of mental images, Piaget (1923) and Luquet (1979) identified three and
five stages of development respectively. Krampen (1991) combined Piaget and Luquet’s phases
and created the following four developmental sequence:
1) Synthetic Incapacity or Scribbling (ages 1-3)
2) Fortuitous and Failed Realism (ages 3-5)
3) Intellectual Realism (ages 5-8)
4) Visual Realism (ages 8-12) (p. 40)
Consequently, children ages 7-9 who have a tendency to draw at the Intellectual Realism Stage,
draw individuals according to their internal mental models, which match the subject whether the
person is posing or not. Krampen (1991) reveals details of what is in a child’s mind and how it
translates into the visual language when drawing in the Intellectual Realistic Stage,
Intellectual realism refers to the fact that children draw everything they know even if it is not
actually visible from a certain vantage point…. Proximities are correctly rendered: thus arms are
attached to the trunk and eyes are placed side by side in the head (even if this is shown in profile).
Order is present in the drawings of landscapes or houses, although no yet in accordance to
coordinates (e.g., the chimney is attached to the house perpendicularly to the roof). Surrounding
and enclosure become so important that in many instances the inside of an object (e.g., a house) is
shown together with its surrounding outline in so-called transparencies. Some projective and
metrical relationships of space are shown, but in an incoherent fashion. Since there is no
coordination of viewpoints or perspectives, a single drawing in this phase may present many
different viewpoints at once (pp. 39-40).
Lange Kuttner, Christine and Reith, Emiel.
Lange-Kuttner and Reith, (1995) base their levels of development on the changing
“conceptualization of objects and space” by examining three components: objects and emptiness,
graphic configurations, and measurement. They claim that as a child develops he acquires
reversibility of thought allowing him “to deconstruct and construct, to structure and measure both
real objects and graphic forms” (p. 87). The subject becomes able to modify “simple geometric
forms into irregular complex shapes, or the supplementation of human figures with spatial axes
systems” (p. 91). Their developmental sequence progresses from global to articulated drawings.
Lowenfeld, Viktor.
Lowenfeld (1939) revisited Piaget’s theory of development in children’s drawings and proposed a
framework for analysis organized thematically. Lowenfeld affirmed that art was an indicator of
children’s thinking as well as the expression of their emotions. He demonstrated that successive
drawings on a theme reflected development.
“Lowenfeld believed that children’s growth through art was analogous to the process of organizing
thoughts and the development of cognitive abilities (Malchiodi, 1998, p. 66).
He looked at creativity in subject matter from the point of view of expression of form, space, and
color, based on the premise that “all artistic ability has its roots in universal human dispositions. It
is determined only by the exceptional intensification of the abilities of a certain individual sense” (p.
5).
Lowenfeld (1939) proposed six stages of development, four of which I use in my study. He
delineated his stages based on observation of children’s use of artistic elements such as color, line,
etc. and their role in the expression of emotional meaning. His six stages of development are:
1. Scribbling Stage (2-4 years)
2. Pre-schematic Stage - First Attempts at representation (4-7 years)
3. Schematic Stage - (7-9 years) Definite form concept.
4. Drawing Realism (9-12 years) Symbolic use of form
5. Pseudo-naturalistic Stage (12 years) Reasoning
6. Period of Decision (p. 27).
Under Lowenfeld’ schematic stage children are able to represent concepts and emotions in their
drawings, performing according to their developmental capacity and their ability to manifest their
insights about their self and others in drawing.
Lowenfeld (1952) emphasized the individual’s ability to express his or her unique characteristics
rather than objective reality (Harris, 1963, p. 38). Lowenfeld said “a child will emphasize and
exaggerate on his drawings those parts of the figure, which have special meaning for him”
(Koppits, 1968, p. 94).
Lowenfeld’s (1939) works in the 30’s and 40’s stressed the need to find links between art and the
psychology of children’s personality development. He said, “What is lacking above all is a
psychological analysis of the correspondence between the development of pictorial art and the
total development of the child” (p. 61).
Lowenfeld’s (1939) framework for analysis of drawings is organized thematically and it identifies
graphic indicators (Di Leo 1973; Golomb, 1972; Lowenfeld, 1939). His developmental sequence is
based on his observation of children’s use of artistic elements (such as color, line, and perspective),
which approximately corresponds to Piaget’s stages of logico-mathematical thinking. It is the lack
of empirical data that propelled Eisner (1967) to criticize Lowenfeld’s scale.
It contains numerous assertions that lack adequate documentation. The stages that are described
are not the result of empirical studies using scientific controls to insure objectivity but insightful,
even if dogmatic conclusions drawn form years of experience working with children (p. 14).
In spite of this criticism, Lowenfeld (1970) successive stages of development for drawing have
been the most widely used for the analysis of elementary school children’s art. This is why I choose
to use it in my study as one of the rating systems to identify levels of complexity in graphic
composition: Pre-Schematic (ages 4-7); Schematic (8-9); Realism (9-12 years), and High Realism
(12-14) (pp. 36-40). These stages may correlate with the developmental disruptions identified by
Fischer (1980).
Although Lowenfeld’s (1939) stages include ways to identify the art of adolescents, the predictable
nature of drawings produced by children older than 12 years of age is less certain (Davis, 1991). In
fact, adolescents may not reach the last stage of artistic development “unless they continue or are
encouraged to make art (Malchiodi, 1998, p. 67).
The adaptation of Fischer’s scale of cognitive development proposed in this study offers a
compelling alternative to Lowenfeld’s scale which is out-dated and does not take into account the
content and meaning of children’s drawings.
Developmental Sequences Of Graphic Representation
In the largest attempt to put together a comprehensive sequence of graphic representation,
Rhoda Kellogg (1966, 1969, 1970) greatly influenced by Goodenough, created a comprehensive
pictorial continuum of the development of children’s drawings of the human figure from a
quantifiable perspective. Her stages of development based purely on the graphic characteristics in
children’s representations were derived from thousands of scribbles and drawings of the human
figures she collected from around the world. Kellogg, however, limited her collection to drawings
from children in the pre-school years.
Goodnow, Jacqueline.
Goodnow (1978) provides descriptions of devices children use in order to show human figures in
movement. She proposes a developmental sequence that includes 15 features, which include
increasing details, dimension, overlaps, manipulation of body axis and bending or flexibility.
Golomb, Clair.
Golomb (1972), author of The Child’s Creation of a Pictorial World, began to look at elementary
school-aged children’s development of graphic representation pointing to their conceptual nature.
Her work focused on “the principles that underlie the graphic mode of visual symbolization” (p. 3).
She observed the common themes and gender preferences that children exhibit at certain ages
and explored the individual and cultural variations on those themes (p. 6).
Golomb (1972) used age and grade levels as ‘convenient ordering devices’ and defined stages of
development as “changes that are largely propelled from within. Stages are determined by the
child’s own and active experimentation with the medium, as well as by the nature of the objects
she wishes to depict” (p. 6).
Golomb contributed to the perspective that competency development is domain specific and that
art is a cognitive activity. In his theory of Multiple Intelligences, Gardner (1983) suggests eight
domains in which the self can develop competencies. With reference to art, Gardner says that it is
an activity of the mind, rather than that of the senses. Students perceive, think, feel and then
express as they produce art. Gardner looked at children’s drawings from a developmental
perspective also organizing them in thematic frames.
Gardner, Howard.
In his book Artful Scribbles, Gardner (1980) argues that children’s drawings may exhibit some of
the same qualities of established artists, and determined that artistic development begins with
scribbles and ends with a search for realism. Gardner proposed a framework for aesthetic
development based on Piaget’ stages. Gardner’s (1980) work focused mostly on development of
aesthetic perception, asking children to report on their ideas about works of art displayed to them.
Willats, John.
Willats (1995, 1997) conducted a study of ways that children represent a table with different
objects on its top. According to his findings, the changing shapes chosen for the representation of
their specific point of view and the foreshortening mark development.
Davis, Jessica.
Davis (Davis, 1991, 1993, 1997; Lawrence-Lightfoot & Davis, 1997) who looked at representational
and non-representational artworks done by children and adults was interested in vehicles and
strategies individuals use to express three specific emotions: happiness, anger and sadness. She
demonstrated the theory of U-shaped development (Davis, 1991, 1997) in terms of expressive and
affective characteristics of the three different emotions. In addition to the expressive qualities of
the collected art work, Davis analyzed the content of the formal artistic properties such overall
expression, overall balance, appropriate use of line and appropriate use of composition. She found
that children’s facility with graphic symbolization most often submerges by age twelve. Children
ages 7-9 work in what is called the “literal stage” and are appropriate candidates for the
production of detailed interactive drawings.
Emphasis for analyzing very young children’s drawings include the study of scribbles (Kellogg,
1966, 1969, 1970; Gardner, 1973, 1994, 1980); as well as representations (Eisner, 1967; Di Leo,
1973, 1977, 1983; Eisner, 1988; Davis, 1991, 1993, 1997; Lawrence-Lightfoot &Davis, 1997;
Goodenough, 1926; Golomb, 1972; Goodnow, 1977; Goodman, 1978; Krampen, 1991; Varon,
1997); introspections (Jaffe, 1979; Lowenfeld, 1939); and retrospectives (Eng, 1931, 1957;
Fischer & Watson, 1980; Varon, 1996, 2000).
Comparative Studies
Many comparative studies emerged and numerous techniques were developed to study children
with specific problems using Goodenough & Harris basic way to categorize and analyze children’s
figure drawings. This technique has been mostly used as diagnostic aids in psychological testing (Di
Leo, 1973, 1977, 1983; Klepesch & Logie, 1982; Koppitz, 1968, 1984; Kramer, 1971, 1993; McNiff,
1992; Peterson & Hardin, 1997).
Comparative studies using control groups include those looking at race, gender, socio-economic or
cultural backgrounds, behavioral, physical, or mental abilities groups, etc. These studies noted the
differences in graphic treatment and occurrences of such treatments by the particular populations
(Eisner, 1967; Klepesch & Logie, 1982; Koppitz, 1968, 1984; Lange-Kuttner & Edelstein, 1995).
In her cross-cultural study, Koppits (1984) found that Argentinean boys and girls drew more
figures in action than American children (p. 72). Koppits did not report explicit details about
children’s linguistic experiences and attributed the difference to children’s cultural experience.
Koppitz, Elizabeth.
Koppits’ (1984) cross-cultural comparative study finds that Argentinean boys and girls draw more
figures in action than American children (p. 72). Koppits attributes this difference to children’s
cultural experience.
Eisner, Elliot.
Eisner (1967) makes reference to the relationship between drawing stages and language
development, and states “…the linguistic system that an individual learns serves to structure his
perception and conception of the world” (Eisner, 1967, p. 34). Eisner’s (1967) report, A
Comparison of the Developmental Drawings Characteristics of Culturally Advantaged and Culturally
Disadvantaged Children, documents that culturally disadvantaged children were lagging behind in
development, “it takes four years for the disadvantaged group to perform at a level that
approaches the culturally advantaged first graders” (p. 72). When he compared developmental
characteristics in the drawings of culturally advantaged and disadvantaged children (from grades
one, three, five and seven), he found that culturally disadvantaged children were lagging behind in
development.
Eisner set up a scale that correlates with cognitive development to assess stages of development
found in children’s art. He compared two groups of students and examined the relationship
between children’s language and artistic development. His scales constitutes of 14 categories
“directly related to the presence or absence of a horizon line or lines and the placement of
morphemes” (p. 35). Consequently, Eisner’s scale is of the same nature as Piaget and Lowenfeld in
that it measures graphic composition or “aesthetic syntax”. If graphic composition measures
aesthetic syntax, I could claim that my proposed visual scale measures “visual semantics”.
Goodenough, Florence.
Goodenough (1926) found that girls represented differentiated dimensions of the human figure
with more details than boys; and boys depicted movement more frequently than girls.
Lange-Kuttner, Christina & Edelstein, Wolfgang.
Lange-Kuttner & Edelstein (1995) conducted a longitudinal study of social factors to graphic
competence and human figure concept. They claim that the transformations of children’s cognitive
development are shaped by their social realities. They found that only drawings of the human
figure were impacted by social variables. “The differences in depiction of volume and movement
varied both as an effect of social class and gender” (p. 163).
Gender differences in drawing do not merely relate to the biological differences between the
sexes… The present study shows that girls were more likely to differentiate the human figures
while boys were more often reduced complexity or depicted movement (p. 169).
Krampen, Martin.
In his comparative study Krampen (1991) looked at cultural differences in the drawings of two
different kinds o buildings. He studied Turkish and German children from urban and rural areas.
He found no difference between the urban children and between the rural ones. Urban children
distinguished more clearly between types of apartment buildings and residential houses than village
children.
In short, research on the development of figure drawing has examined children’s mental and
emotional capacities and development across age, gender, race and cultural experiences. Such
studies have provided in-depth understanding of personality and behavioral aspects of the
development of self in relation to others; and they confirm that children’s drawings are in fact a
reliable measure of development.
The implication of these findings for my study is the confidence that drawings will provide reliable
evidence of children’s cognitive structures, representing their developmental levels, which in turn,
may also be compared. It is also clear that the question still remains as to how cognitive
competency and development in one domain has impact upon another.
Comparison of drawings produced by bilingual and monolingual children, two groups with radically
different cognitive experiences, is unusual. Few studies have investigated the possible effects of
bilingualism on children’s artistic development. Given that linguistic experiences have significant
cognitive impact, the proposed study may provide insight into the effect of bilingualism in children’s
cognitive development.
My study is an attempt to cross the historic boundaries that have existed between different
domains, and is the first one comparing the drawings of monolingual and bilingual children for their
representative characteristics, exploring interaction between graphic composition, language
development, and the visual content.
My intent is to examine drawings of children ages 10-12 because they seek “to make substitute
images of selected aspects of the world about them. However abstract their representations,
children’s purpose [in drawing] is to create graphic equivalent of something perceptually
observable” (Duncum, 2000, p. 44). Often they include scenes of themselves interacting with
others (Davis, 1991; Lowenfeld, 1939; Varon, 2000). I researched frameworks currently
used to identify levels of artistic and cognitive levels of development and explored how these two
perspectives can inform each other.
Bilingualism
There are far more studies that cite the advantages of bilingualism than those that point to its
negative effects.
There are close to 150 empirical studies carried out during the past 30 or so years that have
reported a positive association between additive bilingualism and students’ linguistic, cognitive, or
academic growth (Cummins, 2000, p. 37).
The debate on the effects of bilingualism dates as far back as 1935 when Vygotsky (1962)
suggested that bilingualism alters an individual’s abstract thought processes. Cummins (2001)
discussing the opposite view says, “The Piagetian School has consistently held that the
development of the basic cognitive schemata owes little to language.” (p. 20). Piaget (1967) in
fact, stated that cognitive growth is hardly affected by bilingualism.
The continuous supported acquisition of two languages promotes cognitive, linguistic and
academic growth (Collier, 1987; Cummins, 1981; Hakuta, 1986).
As word meanings and concepts develop. They become increasingly embedded in a system or
network of logical and semantic inter-connections (Cummins, 2000, p. 61).
The process moves swiftly from the beginning level to oral fluency found in natural language.
Academic fluency in two languages (Academic Language Proficiency CALP or (bi-literacy) however,
takes five to seven years. CALP is essential in order to attain full cognitive benefits of bilingualism
(Cummins, 1981).
It reflects the registers of language that children acquire in school and which they need to use
effectively if they are to progress successfully through the grades. The distinction [between natural
and academic language] highlighted the range of cognitive demands and contextual support
involved in particular language tasks or activities (Cummins, 2000, p. 59).
In a recent publication, Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001) found that
bilingual and monolingual children develop at similar rates in biologically controlled mechanisms
related to linguistic acquisition. Bilingual children achieved the same milestones (for example,
number and types of utterances) developmentally the same as monolingual children. Petitto &
Kovelman (2003). found that bilingual children performed better than monolingual children in tasks
that involved abstract reasoning and attention. Petitto (2003) found that bilingual children
performed better than monolingual children in phonological tasks. In addition, through the
analysis of sequential or simultaneous language mixing tasks, they found that bilingualism did not
create confusion.
Negative Effects of Bilingualism
Few studies maintain that bilingualism causes cognitive deficits such as: mental confusion (Saer,
1923), linguistic handicap (Pintner & Keller, 1922), phonic interference (Weinreich, 1957), semi-
lingualism (Skutnabb-Kangas & Toukomaa, 1976) and lower score on verbal ability tests, one of
the most common findings (Ben-Zeev, 1977, Smith, 1933). These studies have been criticized for
their methodology: one such criticism is the lack of matching control groups (Lambert, 1972); and
another is that bilingual participants in these studies had been subjected to subtractive educational
models of bilingualism (Cummins 1976).
Thus, the early findings of negative cognitive effects associated with bilingualism would be
explained by the fact that minority language children in these studies often failed to develop
sufficient high level of proficiency in the school language to benefit fully from the educational
experience (Homel & Palij, 1987, p. 60).
Positive Effects of Bilingualism
Studies putting forward that bilingualism produces general advantages in cognitive abilities are
quite numerous. Findings commonly cite that bilingual children show greater right hemisphere
involvement in their processing language (Albert & Obler, 1978) especially for children, ten years or
older, who are fully bilingual (Genessee, 1978). Stark, Genesee, Lambert and Seitz (1977)
determined that bilinguals developed more laterized verbal processing systems (Homel & Palij,
1987, p.68). Furthermore, bilingualism increases children’s mental flexibility and abstract
reasoning, facilitating their inclination to reorganize cognitive schemes (Ben-Zeev, 1977; Peal &
Lambert, 1962; Petitto, 2003), and promoting divergent thinking (Cummins, 1975).
The most frequently cited study comparing the cognitive performances of monolingual and
bilingual (French-English) students was conducted by Peal and Lambert (1962). They compared
English Speaking Canadians who attended French Immersion Public School in Montreal, Canada
with English Speaking Canadians who only received one period of French a day. Their study was
conducted during emotionally and politically charged times on all sides of the “language” debate,
before French was made the official language of Quebec. They suggested that two languages
result in “mental flexibility, a superiority in concept formation, and a more diversified set of mental
abilities in the sense that the patterns of abilities developed by bilinguals were more diversified”
(Lambert, 1972, p. 152; Lindholm, 1992, p. 206). Peal & Lambert (1962) concluded that bilinguals
performed significantly better than monolinguals on most measures of verbal and non-verbal
intelligence, tasks that required mental manipulation and reorganization of visual patterns, concept
formation, and tasks that required mental and symbolic flexibility (Lambert, 1972, p. 154).
Although Ben-Zeev (1977) makes the case that bilingualism retards vocabulary development, she
agrees that bilinguals devise ways to cope with the complexity of managing two languages,
therefore enhance their cognitive abilities. Ben Zeev argues that in organizing two linguistic
systems, bilinguals develop more analytic approaches, promoting cognitive growth (Been-Zeev,
1977, p. 1010).
After conducting the St. Lambert’s Experiment Study in Eastern Canada, Lambert (1972) suggest
that bilingual children perform better than monolingual children in cognitive tasks due to their
enriched experiences in two languages.
The emergence of an intellectual factor is dependent on the accumulation of experiences… The
bilingual child has been exposed to a wider range of experiences that the monolingual, because his
experience stems from two different [languages and] cultures. This enriched environment may
benefit him on nonverbal tests Higher levels of abstract cognitive processing are developed when
children learns to use two languages, generating symbolic reorganization, agility in concept
formation and improved performed at field independent tasks (pp. 121-40).
According to Baker (1988) the St. Lambert Experiment research project provides some
confirmation that bilingualism is linked to greater field independence. To the contrary, Genesee,
Hammers, Lambert, Monomen, Seitz & Starck (1978) claim that field independence tends to
preclude the acquisition of a second language. In addition, Cummins (1975) found that bilinguals
were superior at tasks that required divergent thinking.
Other frequently declared advantages to bilingualism are greater sensitivity to non-verbal
communications (Skutnaab-Kungas, 1981) and metalinguistic awareness (Bialyosky 1987,1988;
Hakuta &Diaz; 1985; Galambos & Hakuta, 1988; Lasagabaster, 1988; Mohanti, 1978, 1987;
Ricciardelli, 1992, 1993). Hamers and Blanc (2000) emphasize how bilinguals develop special
‘reflective’ skills, which foster other meta-cognitive processes contributing to bilinguals enhanced
learning skills (p. 91). These reflective skills may influence how children perceive and visually
represent events and emotions in their lives (Varon, 2000). Furthermore, Baker mentions studies
where “bilinguals pay attention to the fine details of each social situation” (Baker, 1988, p. 35).
Baker (1988) for example, in reviewing Bain (1975), states that bilinguals had greater sensitivity to
emotional expressions: they were able to better recognize appropriate feelings (Baker, 1988, p.
35).
Recent comparative studies conducted in the United States show how significant the educational
model and levels of bilingual proficiency is to the results. In the right context, two-way immersion
programs for the most part, bilinguals have demonstrated greater academic performances and
more positive cross-cultural attitudes (Cazabon, Lambert & Hall, 1993; Cazabon, Nicoladis, &
Lambert, 1998; Collier, 1987, 1995; Collier & Thomas, 2000; Genesee, 1999; Griego-Jones, 1995;
Lambert, 1977; Lambert & Tucker, 1972; Lambert & Gardner, 1973; Lambert, 1987; Lambert &
Cazabon, 1994; Hakuta, 1986, 1987, 1990, 1999; Skutnabb-Kangas, 1981; Thomas & Collier,
1997, 2002).
The parallel development of two languages is called “additive bilingualism” (Lambert, 1987) as
opposed to “subtractive” bilingualism, in which immigrants learn the language of the dominant
culture at the expense of their heritage language. Gonzalez (1999) affirms “additive bilingualism in
early childhood can accelerate cognitive growth if the child has relatively high proficiency in both
languages” (p. 22). Lindholm (1992) explains, “…additive bilingualism is associated with high levels
of proficiency in two languages, adequate self-esteem, and positive cross cultural attitudes” (p.
197).
Lambert contrasts the subtractive bilingualism of many minority language children with the additive
bilingualism generally achieved by children whose L1 [primary language] is dominant and
prestigious and in who danger of replacement by L2 [second language]. This is the situation of
Anglophone children in French immersion programs… The majority of studies reporting cognitive
advantages associated with bilingualism have been attained an additive form of bilingualism, that is,
relatively high levels of proficiency in both languages (Homel & Palij, 1987, p. 59).
Bilingual Education in the United States
In the US, the Bilingual Education Reform has inspired the implementation of maintenance and two
way immersion bilingual programs, providing more effective than other program models such as
structured immersion or transitional bilingual (TBE). In spite of current research findings pointing
to the most effective ways to educate bilingual children, the debate on the effectiveness of bilingual
education continues to be in the forefront of the political domain. In 1971, the Transitional
Bilingual Education Law was passed making bilingual education mandatory in Massachusetts. Any
district, where there are 20 or more Limited English speaking students from the same linguistic
background, the school system is required to offer native language instruction to those students
for up to three years.
Bilingual education began as a means of helping immigrant students understand school-work while
acquiring English, and maintain their linguistic and cultural identities. The law sought to provide
Limited English Proficient (LEP) students with greater opportunity for academic success than those
students who are immersed in English-only programs (Chapter 71 A, Transitional Bilingual
Education Law). As opposed to maintenance and two-way immersion programs, both TBE and
structured immersion do not target the development of students’ native language. Rather, they
promote the fast assimilation of students into all English programs, which in effect, devalue the
students’ native language.
Indeed, several reports (Porter, 1996; Rossell & Baker, 1996) have been critical of the bilingual
educational model called “Transitional Bilingual Programs”. This type of program was designed to
provide native language support while immigrant children learned English. In this educational
model, students gradually withdraw from all native language instruction, showing signs of what
Lambert (1972) called subtractive bilingualism. These critical reports subsequently do not refer to
detrimental effects of bilingualism as a condition, but rather attack the way children are treated in
the US when they enter transitional bilingual programs. “Teaching a LEP [Limited English
Proficient] student in the native language is at least marginally detrimental to his or her overall
education and acquisition of English” (Rossell & Baker, 1996, p. 2). These authors claim that
students enrolled in transitional bilingual programs will consequently fall behind in all academic
subjects.
In California (Proposition 227, 1999), Arizona (Proposition 203, 2000) and very recently
Massachusetts (English for Children Initiative), bilingual education has already been abolished
through state referendums and replaced with structured immersion programs. One exception, so
far, is in Colorado where the citizens voted to protect their bilingual programs. In spite of the
political trends, research on bilingual education suggests that two-way bilingual immersion
education is effective in educating balanced bilingual children because it is based on an additive
model of bilingualism. In the two-way immersion model of implementation, students from diverse
linguistic backgrounds develop and master the content areas in two languages simultaneously.
Two-Way Bilingual Immersion Programs
Two-way or dual-immersion programs were based on the Canadian model of immersion programs
for foreign language instruction, yet they also integrate elements of the developmental bilingual
model for native language instruction. In the United States dual-immersion programs have
provided equal opportunity for Limited English Proficiency (LEP) students and Native English
Speakers from diverse linguistic backgrounds to be integrated for all subjects and each group to
become proficient bilinguals.
Also during the last three decades, the dual immersion bilingual context has been fertile ground for
research studies, and consensus exists as to their benefits in academic, social, and cognitive areas.
In fact, students in two-way programs have been shown to develop positive cross-cultural
attitudes and they out-perform their monolingual counterparts in academic subjects such as
language arts and mathematics (Cazabon, Lambert & Hall, 1993; Cazabon, Nicoladis, & Lambert,
1998; Christian, 1994; Collier & Thomas, 2000; Genessee, 1987, 1994, 1999; Howard, 2000;
Lambert & Cazabon, 1994; Lindholm, 1992, 1992; McLaughlin, 1995; Thomas & Collier, 1997,
2002).
Collier and Thomas (2002) conducted a five-year longitudinal study that examined academic
achievement of 210,054 students (80 primary languages) from eight different models of bilingual
program including two-way immersion. The demographics in three of the five research sites
focused on Spanish speakers, the largest group of school age population (75% of US linguistic
minority students). The study used standardized tests used by the five urban and rural research
sites across the United States. In general, Collier & Thomas found that bilingual students out-
performed their monolingual counterparts in all academic subjects at the elementary level.
Bilingually school students outperform comparable monolingually school students in academic
achievement in all subjects, after 4 to 7 years of dual language schooling (p. 9).
Native English speakers in two-way bilingual immersion programs maintained their English, added a
second language to their knowledge base and achieved well above the 50th percentile in all
subjects areas on norm-referenced tests in English (p.7).
Thomas and Collier found that Native Spanish Speakers also benefited from the two-way
immersion program model, achieving a 71st percentile after two years in the US.
As of today, no studies investigating the effects of bilingualism have employed the analysis of
students’ drawings to examine cognitive outcomes. Since drawings manifest students’ abstract
thinking as well as their internal representations of social behaviors, and “during the elementary
school years, boys and girls can express their thoughts and feelings often better in visual images
than in words” (Koppits, 1984 p. 2), for my dissertation, I will analyze drawings to examine
students’ cognitive aptitudes.
Hypothesis
I anticipated that bilingual speakers (ages 10-12) would reflect higher levels of cognitive complexity
than those produced by their monolingual counterparts in their graphic compositions, visual
representations of interactions and their verbal descriptions of portrayed interactions.
I anticipated that bilingual speakers (ages 10-12) would reflect higher levels of cognitive complexity
than those produced by their monolingual counterparts in their graphic compositions, visual
representations of interactions and their verbal descriptions of portrayed interactions. My results
may coincide with Vygotsky (1962) in his idea that knowing two language-systems influences the
cognitive condition, suggesting that enhancing development of one domain determines
development in another. This view considers that social and cultural influences (language in
particular) determine and precede cognitive capacities. This view considers that teaching propels
or accelerates development; thus development is context specific.
Vygostky (1935, p.14) suggested that when the application of sound pedagogical principles
ensured that each language had an independent sphere of influence, bilingualism could orient the
child towards more abstract thought processes “from the prison of concrete language forms and
phenomena (Cummins in Homel & Pajil, 1987, pp. 58-59).
Conversely, my results may support Piaget’s claim that cognitive development is concurrent or
precedes linguistic development. Both cognitive and linguistic developments are linked to the early
development of intelligence and to a child’s capacity for symbolic representation (mental
structure). This view postulates that development is universal and not culture specific and that it is
achieved through the manipulation of objects (not necessarily through the development of
language).
Piaget argues that these stages in cognitive development are neither biologically programmed nor
given by the environment, but instead constructed or created by children as they actively attempt
to understand the world (Feldman, 1999, p. 59).
In short, even though both Vygotsky and Piaget agree that environmental influences determine
children’s cognitive development, Piaget attributes development to the interaction between a child’
s innate capacity and his or her experience (assimilation and accommodation), whereas Vygotsky
considers that language exists in thought and that with systematic cooperative interactions
development can be produced regardless of a child’s innate condition.
Both monolingualism and bilingualism are linguistic conditions. One language may exist separate
and may function as a concrete, linear, successive, and consecutive system. Two languages may
exist simultaneously and are always in relation to one another, possibly producing additional
reflections, connections and abstractions, which in turn may function as concurrent and
synchronized abstract systems. Thus, my overarching question: How do these possible different
linguistic processes (consecutive vs. concurrent) impact children’s cognitive structures? If so, how
are they captured in children’s drawings?
In fact, Been Zeev (1977) found that the precocious appearance of categorical associations in
bilingual children developed earlier orientation towards “reorganization”. Bilinguals performed
better at “symbol substitution” tasks, developed an increased ability to analyze non-verbal
structures, and exhibited more mature behaviors than their monolingual counterparts, which
suggests bilingualism promotes precocious development (p. 1015).
When individuals become bilingual, they develop the ability to go from one language to the other,
transfer language skills from one language to the other, and arrive at levels of abstraction that
monolingual individuals are less likely to.
Along the same lines, Lambert (1987) and Bialystock (1999) suggest that bilingual children develop
“some sort of a tri-dimensional view of language, a stereo-scopic perception to which the
monolingual child does not access” (Hamers & Blanc, 2000, p. 85). This relates to Petitto’s finding
that the learning of two languages, in spite of their ability to switch from one language to the
other, does not confuse bilingual children.
The ability to know and to use two or more languages fluently and appropriately reflects the
remarkable functioning of a cognitive system that maintains separation among languages being
used while at the same time allows a free and easy interchange among them (Homel and Pajil 1987,
p. 131).
Research Questions
1. How do the cognitive structures of monolingual and bilingual speakers compare when
performing in three distinct modes of representation: graphic composition, visual content, and
verbal narrative?
2. How do the cognitive structures of monolingual and bilingual speakers compare when
examining their representations of three distinct emotions (happy, sad and angry)?
3. How do monolingual and bilingual speakers (grouped in three distinct ways) compare in
degrees of cognitive complexity as manifested in the graphic composition of their drawings?
4) How do they compare in the degrees of cognitive complexity manifested in the visual
representation of interaction in their drawings?
5) How do they compare in the degrees of cognitive complexity manifested through the verbal
descriptions of depicted interaction as measured by skills theory?
Methodology
Setting
The study was conducted in an urban school (K-8) in Massachusetts with approximately 550
students, 61 % of which participated in a free or reduced lunch program. The setting was ideal
because two different instructional programs, monolingual and two-way bilingual immersion, were
housed in one building. There were two bilingual classrooms and one monolingual classroom in
each of the participating fifth and sixth grades.
Subjects
Subjects attended the same public school in Cambridge, MA. Monolingual (42.39%) and bilingual
(57.61%) speakers were enrolled (for the most part) in two distinct educational programs:
monolingual (English only) program and a bilingual (Spanish/English) dual immersion program.
There were 9 Spanish monolingual students in the bilingual setting who were recent immigrants,
and 9 bilingual (Spanish/English or Portuguese/English) students enrolled in the monolingual
program whose home-language is other than English.
Data was collected from a total of 92 students, 39 of which were enrolled in the monolingual
program and 53 in the bilingual program. Forty-two were girls (45.65%) and 50 were boys
(54.65%). Half of the students (50.55%) participating in the study were eligible for free lunch, 5
students (5.49%) received reduced lunch, and 40 students (43.96%) pay for lunch. Half of the
students (50%) were Hispanic, a third were Caucasian (32.61%) and the rest were African
American (14.13%), Native American (2.17), and Asian (1.9%).
My final sample size is not large, which leads me to recognize that my findings are not generalized
to student-populations at large. I was, however, able to carry out a number of convincing analyses
that reveal hints or trends and that contribute to our understanding of the development of skills
among monolingual and bilingual students as I carried out an in-depth investigation of three
drawings per student. In all, 92 students produced 276 drawings.
Raven Progressive Matrices Test
To insure that groups (monolingual and bilingual) are similar in their basic levels of intelligence, I
obtained scores from their performance on the Raven Progressive Matrices (Raven, 1986) and
groups were matched. An expert who administers and scores this test for the school system
provided me with students’ scores. I used their scores in my analyses to be sure that groups are
not different in basic level of intelligence.
Non-verbal measures of intelligence assure that groups are similar in kind.
In order to isolate the effects of linguality, bilingual and unilingual groups should be matched on
any personal or background characteristic which might contribute to performance on the
dependent variables (Cummins, 2001, p. 33).
The Raven Progressive Matrices have been widely used as a non-verbal measure of intelligence.
Before Lambert’s St. Lambert Experiment (Lambert, 1972), The National Society for the Study of
Education (1928; 1940) published a series of studies that measure the impact of bilingualism on
intelligence citing positive effects (Goodenough, 1928, 1949; Terman 1928, 1949; Stoddard &
Wellman, 1928, 1949).
Distribution of Students Participating in the Study
The breakdown of students by grade level indicates that in fifth grade (ages 10-11) there are a
total of 42 students: (18 (19.57%) in monolingual classes and 24 (26.09%) in bilingual classes).
In the sixth grade (ages 11-13) there are a total of 50 students: 21 students (22.83%) in
monolingual and 29 students (31.52%) in bilingual classes).
Table 1
Number of Students per grade and per educational program
Students 5th Grade 6th grade Total
Monolingual Program 18 21 39
Students in Two-Way Bilingual Program 24 29 53
Total 42 50 92
When grouping students according to their linguistic abilities, there were 53 (57.61%) bilingual
speakers and 39 (42.39%) monolingual speakers.
Table 1.1
Distribution of students according to their linguistic ability
Grade English Monolingual in Monolingual
Program (MM) Spanish Monolingual in Two-Way Bilingual Program (MB) Bilingual in
Two-Way Bilingual Program (BB) Bilingual in Monolingual Program (BM)
5th 12 (13.04%) 4 (4.35%) 20 (21.74%) 6 (6.52%)
6th 18 (19.57%) 5 (5.44%) 24 (26%) 3 (3.26%)
Total 30 (32.61%) 9 (9.78%) 44 (47.83%) 9 (9.78%)
39 (42.39%) 53 (57.61%)
Procedures
Data Collection
I collected three drawings per student in three consecutive sessions, from a total of 92 students
with 2 different linguistic abilities, in either a monolingual or a two-way bilingual immersion
programs. During each session, students worked in their classroom settings and produced one
drawing per session. For practical reasons, since there were two bilingual classrooms per
monolingual one for each grade level. I grouped students from bilingual classrooms of the same
grade level for data collection purposes. Students worked individually, sitting in small groups, while
classroom-teachers were present. Introductory prompts were given orally and most always in
English. If needed, students translated for each other.
All data collection sessions and follow up interviews were observed for bias by a constant outside
witness whose observations were done according to protocol (Appendix B). This individual
was unfamiliar with the school and the distinct programs where data was collected.
To scaffold students’ efforts (Vygotsky, 1978), I showed book-illustrations of relationships and
specific emotions (happy, sad or angry) and encouraged students to share what they saw. All
students looked at the same illustrations before I introduced the task to the whole class at once.
Students each drew a picture after I read the “happy” prompt the first session; “sad” prompt the
second session and “angry” the third session (Appendix C). The logic used for ordering the tasks
was that “happy” was best to start with, to set the tone; “sad” was second because I did not want
it to be last; and “angry” was third by default.
I provided students with pencils, paper, and asked them to indicate their names, grade level, and
classroom number on the back of their papers. Children worked until they considered their
drawings finished, which typically took approximately 20 minutes. Students held on to their
drawings and waited to be called for a follow-up interview.
Interviews
Immediately following each drawing session, I conducted brief individual interviews (approximately
3-10 minutes per student) in an adjacent library. Children’s descriptions of their drawings using
questions (Appendix D) designed to capture their intentions, thoughts, and descriptions of the
relationships, places, and interactions portrayed in each drawing were tape-recorded. I
encouraged students to give as much information about their drawings as they could.
To ensure the most eloquent narratives all Spanish monolingual students in the bilingual program
were interviewed in Spanish; and all bilingual students in the monolingual classrooms, as well as all
other participants, were interviewed in English.
Scoring Procedures
Inter-rater Reliability.
An independent rater and I scored 20 % of all drawings for their graphic composition, 20 % of all
transcriptions for their verbal narrative, and 40 % of all drawings for their visual content. This was
done to determine if different raters could reliably use the scales.
The independent rater for the drawings was a veteran art teacher with a Bachelor’s in Fine Arts
and Psychology from Emmanuel College and a Masters Degree in Expressive Therapies from Lesley
University.
The independent rater for the verbal narratives was former student of Kurt Fischer who graduated
with a Master’s Degree from Harvard University. For the verbal accounts of the nature of
interaction in students’ drawings, ratings were done based on the content in students’ answers
(from transcribed tapes) using Fischer’s criteria for the various levels of complexity (Appendix A).
Both raters demonstrated great understanding of the theories and scales that they used and had
no previous knowledge of my study or the students who participated in it. During the process,
raters had no indication of students’ linguistic classification, grade level, or program enrollment.
For practical purposes, students were assigned identification numbers and data sets and scoring
spreadsheets were color-coded.
The levels of agreement between raters were very high:
- for the graphic composition we agreed on 48 out of 53 drawings = 91%
- for the verbal narratives we agreed on 50 out of 53 drawings = 94 %
- for the visual content we agreed on 81 out of 91 = 89%
These results indicated that the independent raters could continue finishing the rest of the scoring
on their own.
Scoring of Drawings by Independent Raters.
Each student received three scores per drawing, one in each category: graphic composition, verbal
description of interaction, and visual representation (a total of nine scores for three drawings –
three for each category). Values were pre-designated for each category (Table 2) and criteria for
each category was determined
Table 2 –
Values for Scoring Drawings in Three Categories
Category 1 Category 2 Category 3
GRAPHIC……………….. VERBAL …………………VISUAL
Based on Piaget Based on Fischer Proposed by
Varon
and Lowenfeld (1947) (1980) (2002)
Golomb (1992) Di Lio (1977)
Scores were recorded on color-coded scoring sheets. The scales for graphic composition and
verbal description of interactions had been widely used to measure artistic and cognitive
development. The scale for visual representation was derived from Fischer (1980) and was my
proposed measure to identify cognitive development through drawings. Values in each category
were linked to specific criteria (Appendix A) representing progressive levels of cognitive
development (from simple to complex) from three different perspectives:
1) Graphic composition based on traditional artistic development by looking at the
arrangement and perspective of visual items on the page as identified by Lowenfeld (1939). The
scale (Table 2.1) consisted of numbers from 1 to 4, with 4 capturing higher levels of graphic
composition.
Table 2.1 –
Criteria for Scoring the Graphic Composition Category:
STAGE DESCRIPTION VALUE
Pre-Schematic:
1. Simple: Floating figures or attached to a base line; single point of view with action related to the
self only.
Random collection of objects or figures that do not seem to relate to each other on the page;
- No depth or overlap;
- Eye level view: without any horizon line; or
- Single line
or
- Perpendicular Lines relationships are used to depict space:
- Side or Front view of same object limited to two faces of object or agent, though front face
preferred;
- Transparencies or line Intersections are present 1
Visual Realism
2. Multiple; Figures relating to more than one linear axis. Objects or individuals are arranged
representing the personal perspective of self.
- Children are able to use more realistic proportions of figures and portray them in profile
- Two separate ground lines or combining birds eye view
- Children use fold over perspective - overlaps
- Story line and time sequences are depicted in representation
2
Realism /Literal
3. Sequential: sequence of time and space defined in frames or boxes creating a narrative yet still
from the perspective of self.
- Children are able to depict differentiation between self and others each with own characteristics –
true to appearance;
- Cross gender characteristics
- Detail oriented drawing
- Children are able to depict figures in motion
- Figures are represented only part view (Lange-Kuttner & Wolfgang Edelstein p.71)
- Cartoon-like at times
- Parallelogram perspectives is used - frontal parallel plane 3
Pseudo-Naturalism
4. Global: perspective determined by global composition, combining different points of view, not
limited to the perspective of self.
Children are able to draw the idealized human form
Children are able to use converging perspectives: combining vertical and oblique projections –
Children pay attention to mood and often representations are personal statements.
Children pay attention to greater detail and or abstractions depicting cause and effect/
interdependence of relations.
Children are able to represent three dimensions 4
Coordination of Abstractions
Children are able to create groups of concepts
Simple symbol relationship
2) Verbal description of depicted interaction by looking at the nature of roles and relationships
as evidence of the complexity level as postulated by Fischer (1980, 1993). The scale here ran from
1 to 5, with 5 representing higher levels of complexity (Table 2.2).
3) Visual representation of self in relationship with others by examining the nature of
portrayed relationships and objects in the page rather than composition of graphic elements of
drawings. Levels (Table 2.2) were inferred from Fischer’s cognitive levels and adapted to create an
alternate way of classifying artistic development. This will be scored using the same 1-5 scale of
complexity.
Table 2.2 - Criteria for Scoring the Verbal & Visual Representation Category
Fischer’s Levels – Verbal Relations Score Varon’s Levels - Visual
I- Single Representations
(Rp1)
(Descriptions of criteria provided by Fischer 2003 in e-mail correspondence) No relation
1 Discrete – Distinct Image
Child draws one or more figures or objects without evidence of any relationships or specific location
II- Representational Mappings
(Rp 2)
Involves mapping of two or several items focusing on one aspect of each
Reciprocal
Sequential
2 Coordinated – Sequential Image
Child draws two or more figures/objects in a direct, simple relationship either collaborative or
complementary. Arrangement of figures and objects on the page are not unified or
interconnected.
Child draws figures involved in the same/similar action with little or no differentiation between
them. Child draws figures or object in a generalized context.
III Representational Systems
(Rp 3) Involves relating several aspects of each item in a system Causal
Coordinated
Temporal
Instrumental 3 Integrated Image
Child draws two or more figures/objects that are arranged in an integrated system of concrete
relationships, - At times figures or objects are portrayed in synchronized action/s that produce
reactions within that system. The cause, reason, and context (specific place) for actions are explicit.
Child draws evidence of a temporal relationship.
IV– Single Abstractions
(Rp 4 or Ab 1)
Involves coordinating several concrete inter-relationships (systems) in terms of an abstract or
intangible characteristic
Intangible
Conflictive 4 Figurative – Symbolic image
Child draws one or more concrete figures/objects forming relationships that produce intangible
meaning. These relationships have the clear intent to express implicit characteristic.
Child draws one or more figures/objects involved in differentiated systems of relationships that
result in the portrayal of conflictive or distinct values or attributes.
Child uses universal symbols to express abstract concepts.
V – Abstract Mapping
(Ab 2)
Involves relating two abstracts characteristic with each other Deduction
5 Logical Symbolic image
Child draws one or more figures/objects that are portrayed through symbols expressing a desired
outcome/idea. The image carries clear motivation in the use of symbols - or symbolic objects - to
depict what self intends.
Child combines sets of symbols to convey a meaning beyond the symbols themselves.
For all scoring, two individuals (an independent rater and me) scored and coded at least 20 % of
the data set separately. Where differences in scores appeared between raters, discussions were
held until inter-rater agreement was reached. The independent rater provided the scores used for
the final analysis.
Analysis
To examine monolingual and bilingual students’ drawings, I designed an integrated task that
included three components of what constitutes a representation according to Freeman (1980): a
definable object, definition of a relationship, and the study of action. In the analysis I looked at the
mean scores of group performances in each of three categories (graphic, verbal, and visual) and
translated them into percentages because the scales used to assign scores in each category were
not the same. In order to find the corresponding percent for each category, I divided the score by
the number of steps on each scale, (i.e. all scores for the graphic composition were divided by four;
all scores for the visual and verbal categories were divided by five). Once scores were comparable,
I examined the data by grouping students in three different configurations looking for changes in
their representational structures and the organization of the represented relationships. The
grouping of students followed three different configurations:
1) Monolingual and bilingual speakers;
2) Four distinct sub-groups of monolingual and bilingual speakers grouped according to their
linguistic backgrounds and educational experiences; and
3) Students enrolled in two distinct educational programs (English Monolingual and Two-Way
Spanish/English Bilingual Immersion).
For each configuration of students, I organized the data according to:
a) Three representational modes: graphic, visual, and verbal; and
b) Three distinct emotions: happy, sad, and angry.
Answering Research Questions
My main research questions focuses on the differences in the degrees of cognitive complexity
between the monolingual and bilingual speakers and the differences between students enrolled in a
monolingual English program and a two-way bilingual immersion program. I analyzed three modes
of representation and the representation of three emotions.
1.) As reflected by the three modes of representation together;
2.) As reflected by the representation of three emotions together;
3.) As reflected in the graphic composition of students’ drawings;
4.) As represented in the relationships of the visual content depicted in students’ drawings;
and
5.) As reflected in students’ verbal descriptions of the interactions in their drawings.
I addressed the research questions by carrying out a series of regression analyses, using separate
analyses for each of the main outcomes. The scores from three drawings and corresponding verbal
narratives (for happy, sad, and angry) were analyzed separately and combined. I also combined
the scores from the three different modes of representation: graphic, visual, and verbal. The main
question predictor variable was bilingual status (coded 0 = monolingual program / 1= bilingual
program); and the other predictors were student linguistic abilities (coded 0= English-only / 1 =
Other than English (mostly Spanish or Portuguese + English); grade level (coded 0 = fifth grade /
1= sixth grade).
I examined if the effects of bilingual status or program enrollment would be significant to either the
mode of expression and the emotional content of students’ representations. Thus, I investigated
possible interactions of linguistic and educational status, home language, and emotional content of
drawing. Significant interaction effects would indicate that certain subgroups of students
(Appendix F) perform differently than others.
Finally, I examined the differences between the three scales (different ways to score drawings) and
the relationships between the emotions represented in each drawing. This helped me to see
whether students are relatively consistent in their scores on graphic compositions, visual
representations, and verbal descriptions of interaction, or whether there are notable differences
across the emotions.
Findings
Overall, there were no significant statistical differences across three of the four compared groups.
It is interesting to note that the Spanish monolingual group who scored statistically lower in the
Raven Progressive Matrices showed the few instances where there was a statistical difference.
Comparison Of Students’ Basic Level Of Intelligence
In comparison of the basic level of intelligence using the Raven Progressive Matrices the differences
were statistically significant in all three comparisons: a) across four groups of students, (F=7.71,
p= 0.0067); b) across monolingual and bilingual speakers (F=4.31, p=0.04); and c) across
educational programs (F=4.04, p=0.047).
Pair-wise comparisons indicated that there was a consistent difference between the Spanish
monolingual speakers in the bilingual program (one sub-group) and everyone else. The Spanish
monolingual speakers in the bilingual program scored the lowest (M=45.55%, SD = 36.43) of the
four sub groups. Their score is statistically significantly different than the bilingual speakers in the
bilingual programs (p=0.0007), than the bilingual speakers in the monolingual program (p=0.
0246), and than the monolingual speakers in the monolingual program (p=0.0013). (See table 3.)
Once I segregate the scores of the Spanish monolingual students, all other comparisons of the
Raven scores do not reveal significant statistical differences. Bilingual speakers in the bilingual
program (M=8.25, SD = 29.50) and English monolingual speakers in the monolingual program (M=
8.2, SD= 27.46) scored practically the same with no statistically significant difference between
them (p= 0.9418). Bilingual speakers in the monolingual program (M=76.66%, SD =
20.61) scored in third place, also with no statistically significant difference between them and
bilingual speakers in the bilingual program (p= 0.5821) and the English monolingual speakers in the
monolingual program (p= 0.6281).
The Spanish monolingual students are, for the most part, recent immigrants to the United States
who did not receive comparable contextual and environmental influences while growing up
(Appendix F). Results are presented in table 3.
Table 3
P-Values when Comparing Raven Progressive Matrices Across Four Groups
Bilingual Speakers In Bilingual Program
(BB) Bilingual Speakers In Monolingual Program
(BM) Spanish Monolingual Speakers In
Bilingual Program (MB) English Monolingual Speakers in Monolingual Program (MM)
BB - 0.0007 0.5821 0.9418
BM - - 0.0246 0.0013
MB - - - 0.6281
MM - - - -
The above-mentioned differences between the sub-group of Spanish monolingual students and
the rest affect all comparisons in the various configurations. When comparing the Raven scores of
students in the monolingual (M=80.76%, SD = 25.89) and bilingual (M=76.22%, SD = 33.46)
programs, the difference is statistically significant (F=4.04, p=0.0475). In addition, the difference
in Raven scores between the monolingual (M=73.58%, SD = 33.12) and bilingual (M=81.50%,
SD= 28.10) speakers, which also includes the sub-groups, is statistically significant (F=4.31. p=0.
0408).
A - Aggregated Scores for Three Expressive Modes of Representation and the Representations of
Three Emotions
When aggregating scores of all students and examining the differences across the three modes of
representation (Table 3.1 and Appendix G), the contrast shows that graphic composition
generated the highest score (M=57.78%, SD=15.11), followed by visual representation (M=47.
68%, SD=12.31), and then by verbal narrative (M=39.92%, SD=11.06).
There were overall differences in the levels of complexity ratings across the three modes (F (2,182)
= 73.1, p= <0001). Results are shown in table 3.1. The post-hoc contrasts here revealed that the
ratings for graphic were higher than those for verbal (p=.0001) and visual (p-.0001) and the
difference between verbal and visual was also significant (p=.0001).
When aggregating scores of students in their representations of events associated with three
emotions (Table 3.1 and Appendix H) and examining their differences across emotions, “anger”
produced the highest score (M=50.34%, SD=14.28) which reflects the highest level of complexity,
followed by “sad” (M=49.64%, SD=14.40) in second place and “happy” which got the lowest score
(M=45.29%, SD=11.76).
When looking at the overall performance of all students in their levels of complexity across the
three different emotions (happy, sad and angry), I found that there was a statistically significant
difference (F (2,178)= 5.62, p = 0.004). Results are shown in table 3.1. The post-hoc contrast
indicated that the mean score for happy was significantly lower than the means for sad (p = 0.004)
and angry (p = 0.016). The means for sad and angry were not statistically different.
It seems interesting to note that I found a strong correlation between the visual and graphic
modes (r = .81) and weak correlations between the visual and verbal modes (r = .18); and the
verbal and graphic modes (r = .21). This indicates that it is important to assess children using
different modes if we expected to construct a fuller picture of their cognitive abilities.
This is evidence that the Varon scales, derived from Fischer’s scales for measuring cognitive levels
through behavior and thinking skills, are successful in measuring cognitive development through
drawing. The Varon scales were therefore developmentally appropriate.
The verbal mode and happy emotions produced the lowest levels of complexity, contrary to
graphic representations and anger, which reached the highest levels of complexity. Results are
presented in table 3.1.
Table 3.1
Average Ratings found in Modes of Representation and Representations of Emotions
Level of Complexity Modes of Representation
(F (2,182) = 73.1, p= <0001) Representation of Emotion
(F (2,178)=5.62, p =0.004)
Highest levels Graphic: M= 57.78%, SD=15.71 Angry: M= 50.34%, SD=14.28
Middle levels Visual: M= 47.68%, SD=12.31 Sad: M= 49.64%, SD=14.40
Lowest levels Verbal: M= 39.92%, SD=11.06 Happy: M= 45.29%, SD=11.76
B - Comparison of Monolingual and Bilingual Speakers’ Performance In Three Representational
Modes And Three Emotions (See Table 4 & Appendix I)
When comparing Monolingual and Bilingual Speakers’ aggregated scores in the three modes of
representational systems, F-Tests results indicate that there is no statistical difference between the
two groups (F=1.41, p=0.2381). Bilingual speakers scored only very slightly higher (49.17%, SD=
9.63) than monolingual speakers (47.50%, SD=11.21).
Additionally, when comparing monolingual and bilingual speakers’ aggregated scores in the three
representations of emotions, F-Tests results indicate that there is no statistically significant
difference between the two groups (F=1.43, p=0.2356). Bilingual speakers scored slightly higher
(49.12 %, SD= 9.64) than monolingual speakers (47.48%, SD=11.22).
Table 4
Monolingual and Bilingual Speakers’ Average Rating in all Three Modes of Representation and
Three Emotions
Monolingual Speakers
N = 39 Bilingual Speakers
N = 53 ANOVA
Variables M SD M SD F p
Modes of Representation 47.50 % 11.21 49.17% 9.63 1.41 0.2381
Representation of Three Emotions 47.48% 11.22 49.12% 9.64 1.43
0.2356
B.1- Monolingual And Bilingual Speakers’ Performance In Three Modes Of Representation (See
Table 4.1 & Appendix J)
Results in the three distinct representational modes indicate that there is no significant statistical
difference found between monolingual and bilingual speakers.
Graphic Composition. F-Tests results indicate that there is no statistical difference between the
two groups (F=2.29, p=0.1334) when looking at the average scores of monolingual and bilingual
speakers based on their graphic composition (in a scale of 1-4 based on Piaget and Lowenfeld’s
levels of artistic development), bilingual speakers (58.33%, SD=14.52) scored slightly higher
(57.05%, SD=17.36) than monolingual speakers in graphic representations.
Visual Representation. F-Tests results indicate that there is no statistical difference between the
two groups (F=2.84, p=0.0956) when looking at the average scores of students’ drawings based
on the complexity of their visual representations, (in a scale of 1-5 based on Varon’s adaptation of
Kurt Fischer’s levels of cognitive development), bilingual speakers scored higher (M= 49.55%,
SD=11.74) than monolingual speakers (M= 45.12%, SD= 12.74) in the representation of visual
content.
Verbal Narrative. F-Tests results indicate that there is no statistical difference between the two
groups (F=0.44, p=0.5071) when looking at the average scores of students’ verbal narratives (in a
scale of 1-5 based on Kurt Fischer’s levels of cognitive development). Monolingual speakers (M=40.
34%, SD=12.60) scored slightly higher than bilingual speakers (M=39.62%, SD=9.88) in their
verbal narratives.
Table 4.1
Monolingual and Bilingual Speakers’ Average Ratings in Three Representational Modes
Monolingual Speakers
N = 39 Bilingual Speakers
N = 53 ANOVA
Variables M SD M SD F p
Graphic 57.05 17.36 58.33 14.52 2.29 0.1334
Visual 45.12 12.74 49.55 11.74 2.84 0.0956
Verbal 40.34 12.60 39.62 9.88 0.44 0.5071
Total 47.50 11.21 49.17 9.63 1.41 0.2381
B.2. - Monolingual And Bilingual Speakers’ Performance In The Representations Of Three Distinct
Emotions (See Table 4.2 & Appendix K)
Results in the three distinct representations of emotions indicate that there is no significant
statistical difference found between monolingual and bilingual speakers.
Happy. The F-Test results indicate that there is no statistically significant difference in the
mean (F=1.33, p=0.2524) in students’ representation of happy. Monolingual speakers (M=45.40
%, SD = 11.05) and bilingual speakers (M=45.22%, SD = 12.36) scored virtually the same.
Sad. The F-Test results indicate that there is no statistically significant difference in the mean
(F=1.13, p=0.2909) in students’ representation of sad. Bilingual speakers scored higher (M=51.
53%, SD=13.77) than monolingual speakers (M=47.17%, SD=15.01).
Angry. The F-Test results indicate that there is no statistically significant difference in the
mean (F=0.34, p=0.5617) in students’ representation of anger. Bilingual speakers scored higher
(M=50.69 %, SD=13.12) than monolingual speakers (M=48.87 %, SD=15.89).
Table 4.2-
Monolingual and Bilingual Speakers’ Average Ratings for the Representation of Three Emotions
Monolingual Speakers
N = 39 Bilingual Speakers
N= 53 ANOVA
Variables M SD M SD F p
Happy 45.40 11.05 45.22 12.36 1.33 0.2524
Sad 47.17 15.01 51.53 13.77 1.13 0.2909
Angry 49.87 15.89 50.69 13.12 0.34 0.5617
Total 47.48 11.22 49.12 9.64 1.43 0.2356
C. - Comparison of the Performances of Four Groups of Students Based On Their Linguistic
Backgrounds
This comparison examines the data of four distinct groups of students based on their linguistic
backgrounds and enrollment in educational programs:
(BB) Proficient bilingual students enrolled in bilingual two-way immersion program who
come from Bilingual, Spanish, or English Speaking homes (n = 44);
(MM) English monolingual students who are enrolled in a monolingual program who are from
English-only speaking homes (n=30);
(BM) Bilingual students whose home language is other than English who are enrolled in the
monolingual program (n=9), some of whom are international students; and
(MB) Spanish monolingual students who are native Spanish speakers who are enrolled in the
bilingual classrooms (n=9). These students were all recent arrivals to the United States and new to
the school system.
C.1. - Aggregated Scores For Three Representational Modes (See Table 5 & Appendix L)
The differences across four groups are not statistically significant (F=1.53, p=0.2191) when
looking at the average aggregated scores for all three representational modes expression
together. Bilingual students in two-way bilingual programs scored the highest (M= 49.24%, SD=
9.17). English monolingual students in monolingual program (M= 48.96%, SD=11.99) followed.
Bilingual students in monolingual program, ranked in third place (M= 48.82%, SD=12.28), and
Spanish monolingual students in two-way bilingual program scored the lowest (M= 42.65%, SD=
6.43).
These findings are consistent with differences found in the Raven Progressive Matrices
performances. In effect, even though the differences here are not statistically significant, it is
worth noting the difference in performance between Spanish monolingual students in the two-way
bilingual program and all other three groups. These Spanish monolingual students’ lower
performance is consistent with their lower performance in the Raven Progressive Matrices which is
most likely influenced by their underprivileged backgrounds, limited schooling, and restricted
chances to express themselves in two languages in school.
C.2. - Four-Group Performance In Three Distinct Modes of Representation (See Table 5 & Appendix
M)
When data was desegregated by different representational modes, one of the sub-groups
performed consistently lower than the other groups, but the difference was not always statistically
significant.
Graphic Composition. In graphic composition (in a scale of 1-4 based on Piaget and Lowenfeld’s
levels of artistic development), differences across groups were not statistically significant (F=1.03,
p=0.3121). The only statistically significant difference was when comparing the Spanish
monolingual speakers in the bilingual program as compared with the English monolingual (p=0.
0328) and bilingual (p=0.0457) speakers in the monolingual program. The difference is not
statistically significant when comparing the Spanish monolingual speakers in the bilingual program
with the bilingual speakers in the bilingual program (p= 0.0714).
In their graphic compositions, bilingual speakers in the monolingual program scored highest (M=62.
03%, SD=15.65), followed by English monolingual speakers in monolingual classes (M=60.00%,
SD=17.96). In third place were the bilingual speakers in the bilingual program (M=57.57%, SD=14.
35), and last were the Spanish monolingual speakers in the bilingual program (M=47.22%, SD=11.
02).
Visual Representation. There is no statistical difference across the groups (F=0.74, p=0.3926)
when looking at the average percent scores of students’ drawings based on the complexity of their
visual representations, (in a scale of 1-5 based on Varon’s adaptation of Kurt Fischer’s levels of
cognitive development). Bilingual speakers in the bilingual program scored highest (M=49.69%,
SD= 11.01), followed by bilingual speakers in the monolingual program (M=48.88%, SD=15.63),
then by English monolingual speakers in monolingual program (M=46.22%, SD=13.77). Spanish
monolingual students in bilingual program scored the lowest (M=41.48%, SD= 8.01). These
Spanish monolingual speakers in bilingual classrooms are recent immigrants from Spanish Speaking
countries or Puerto Rico and their “slight lag in development” may be due of diverse environmental
context and psycho-socio-economic factors affecting their educational backgrounds.
Verbal Narrative. There are no statistically significant differences across groups (F=1.15, p=0.
2863) when looking at the average scores of students’ verbal narratives (in a scale of 1-5 based on
Kurt Fischer’s levels of cognitive development). English monolingual speakers in monolingual
classrooms (M= 40.66%, SD=13.82) and bilingual speakers in bilingual classrooms (M= 40.45%,
SD=9.55) scored virtually the same in verbal narratives. Spanish monolingual speakers in the
bilingual programs (M=39.25%, SD=7.77) rated in third place (and most likely because their
interviews were held in Spanish) with a slightly better score than the bilingual speakers in
monolingual programs (M= 35.55%, SD=11.05).
The fact that the bilingual speakers in monolingual programs scored lower in verbal narratives than
the other groups, especially when we comparing their scores with the bilingual speakers in the
bilingual program, suggests that the two-way bilingual immersion program would be more a better
education fit for these students’ verbal development. As slight as this difference may be
(approximately 5 points), the impact of total assimilation and subtractive bilingualism prevents
bilingual home speakers from reaching their full cognitive potential. When students share and
preserve their native languages and cultures they seem to reach optimal potential. A study with
much larger number of students would be necessary to look for evidence of this hinted trend.
Table 5 -
Average Ratings of Drawings Across Four Groups and Across Three Modes of Representation
English Monolingual Speakers in Monolingual Program
(MM) Spanish Monolingual Speakers in Bilingual Program
(MB) Bilingual Speakers in Bilingual Program
(BB) Bilingual Speakers in Monolingual Program
(BM)
ANOVA
Variable M SD M SD M SD M SD F p
Graphic 60.00 17.96 47.22 11.02 57.57 14.35 62.03 15.65
1.03 0.3121
Visual 46.22 13.77 41.48 8.01 49.69 11.01 48.88 15.63
0.74 0.3926
Verbal 40.66 13.82 39.25 7.77 40.45 9.55 35.55 11.05
1.15 0.2863
Total 48.96 11.99 42.65 6.43 49.24 9.17 48.82 12.28
1.53 0.2191
C.3. - Aggregated Scores for the Representation of Three Emotions (See Table5.1 & Appendix N)
F-Test results indicate that there are no significant differences across the groups (F=1.55, p=0.
2168) when examining the aggregated scores of the representation of three emotions. The
comparison across four groups reveals that all students, except the Spanish monolinguals in the
two-way bilingual program scored similarly. Bilingual students in the two-way bilingual program
scored highest (M=49.18%, SD=9.18), followed by the English monolingual students in the
monolingual program (M=48.96%, SD=11.99) and bilinguals students in the monolingual program
(M=48.82%, SD=12.28). Spanish monolingual students in the bilingual program scored the lowest
(M=42.56%, SD= 6.43).
C.4. - Four-Group Performance In The Representations Of Three Distinct Emotions (See Table 5.1
& Appendix N)
Examining students’ performances in the different emotions provide us with a detailed illustration
of how the connection between levels of cognitive complexity and emotion.
Happy. F-Test results indicate that there are no significant differences across the groups
(F=0.17, p=0.6810). Bilingual students in the monolingual program (M=49.44%, SD=6.97) scored
the highest when representing a happy occasion, followed by English monolingual students in the
monolingual program (M=47.16%, SD=10.67) and by bilingual students in bilingual program
(M=44.35%, SD=13.08). Spanish monolingual students in bilingual programs scored the lowest
(M=39.53%, SD=10.82).
Sad. F-Test results indicate that there are no significant differences across the groups (F=1.
12, p=0.2737). Bilingual students in the bilingual program (M=52.30%, SD= 12.66) scored the
highest when representing a sad occasion, followed by English monolingual students in the
monolingual program (M=48.11%, SD=16.79) and by bilingual students in the monolingual
program (M=47.96%, SD=18.61). Spanish monolingual students in the bilingual program scored
the lowest (M=44.07%, SD=5.78).
Angry. F-Test results indicate that there are no significant differences across the groups
(F=1.57, p=0.2137). English monolingual students in the monolingual program (M=51.61%,
SD=16.70) scored the highest when representing an angry occasion, followed by bilingual students
in the two-way bilingual program (M=51.02%, SD=12.68) and by bilingual students in the
monolingual program (M=49.07%, SD=15.85). Spanish monolingual students in the two-way
bilingual program scored the lowest (M=44.07%, SD=11.81).
Table 5.1-
Average Ratings Across Four Groups & Across the Representations of Three Emotions
English Monolingual Speakers in Monolingual Program
(MM) Spanish Monolingual Speakers in Bilingual Program
(MB) Bilingual Speakers in Bilingual Program
(BB) Bilingual Speakers in Monolingual Program
(BM)
ANOVA
Variable M SD M SD M SD M SD F p
Happy 47.16 10.67 39.53 10.82 44.35 13.08 49.44 6.97
0.17 0.6810
Sad 48.11 16.79 44.07 5.78 52.30 12.66 47.96 18.61
1.21 0.2737
Angry 51.61 16.70 44.07 11.81 51.02 12.68 49.07 15.85
1.57 0.2137
Total 48.96 11.99 42.56 6.43 49.18 9.18 48.82 12.28
1.55 0.2168
D. – Comparison of Students’ Performances Based on their Enrollment In Two Distinct Educational
Programs (See Table 6 & Appendix O)
This comparison includes two groups:
Bilingual and Spanish monolingual students enrolled in the two-way bilingual immersion
program regardless of their home language are Spanish or English home-speakers who are being
schooled in two languages (n=53). The monolingual students in this program are recent arrivals
from Spanish speaking countries, which have access to entry in this program at any given grade
level; and
English monolingual students in the monolingual program with English-only as their home
language and bilingual home-speakers enrolled in the same English-only program (n=39).
D.1. - Students’ Performances In Three Modes Of Representation
(See Tables 6 & Appendix O)
The F-Test results indicate that there is no statistically significant difference in the means (F=1.18,
p=0.2809) when examining students the overall performance in the three modes of representation
(graphic, visual and verbal). Both students in the monolingual program (M=48.93%, SD=11.89)
and students in the two-way bilingual program (M=48.12%, SD=9.07) scored very similar to each
other. The lower scores of Spanish monolingual students in the two-way bilingual program that
we saw earlier, who are recent arrivals to the United States, contribute to this difference. In
further analysis, we are able to see that only the scores in students’ graphic representations affect
the difference. Results are presented in table 6 (Appendix P)
Graphic. F-test results determine that there is a significant statistical difference between the scores
of students in both programs in the complexity of graphic representations (F=4.44, p=3.79).
Students in the monolingual program (M=60.47%, SD=17.28) scored higher than students in the
two-way bilingual program (M=55.81%, SD=14.30).
Visual. F-test results determine that there is no statistically significant difference between the
scores of both programs in the complexity of graphic representations (F=0.37, p=0.5442).
Students in the two-way bilingual program (M=48.30%, SD=10.95) scored higher than students
in the monolingual program (M=46.83%, SD= 14.05).
Verbal. F-test results determine that there is no statistically significant difference between the
scores of both programs in the complexity of graphic representations (F=0.35, p=0.5541).
Students in the two-way bilingual program (M=40.25%, SD= 9.21) scored only slightly higher than
students in the monolingual program (M=39.48%, SD=13.27). There are no significant statistical
differences.
Table 6 –
Comparison of Two Educational Programs on Three Modes of Representation
Monolingual Program
N= 39 Two-Way Bilingual Program
N= 53 ANOVA
Variables M SD M SD F p
Graphic 60.47 17.28 55.81 14.30 4.44 0.0379
Visual 46.83 14.05 48.30 10.95 0.37 0.5442
Verbal 39.48 13.27 40.25 9.21 0.35 0.5541
Total 48.93 11.89 48.12 9.07 1.18 0.2809
D.2. - Students’ Performances In The Representations Of Three Emotions (See Table 6.1 &
Appendix O)
The F-Test results indicate that there is no statistically significant difference in the means (F=1.24,
p=0.2691) when examining students’ overall performance in students’ drawings associated with
three different emotions. Students in the monolingual program (M=48.93%, SD=11.89) and
students in the two-way bilingual program (M=48.06%, SD=9.07) scored virtually the same. In
further analysis we are able to see that differences in the representation of emotions are minimal.
Results are presented in table 6.1 (Appendix Q).
Happy. The F-Test results indicate that there is a statistically significant difference in the
means (F=4.26, p= 4.19). Students in the monolingual program scored higher (M=47.69%, SD=9.
90) than students in the bilingual program (M=43.53%, SD=12.76).
Sad. The F-Test results indicate that there is no statistically significant difference in the
means (F=0.00, p=0.9684). Students in the bilingual program (M=50.84%, SD=12.11) scored
higher than students in the monolingual program (M=48.07%, SD=16.97).
Angry. The F-Test results indicate that there is no statistically significant difference in the
means (F=0.54, p=0.4626). Students in the monolingual program scored higher (M=51.02%,
SD=16.34) than students in the bilingual program (M=49.84%, SD=12.70).
Table 6.1-
Comparison of Students in Monolingual and Bilingual Programs by the Representation of Emotions
Monolingual Program
N= 39 Two-Way Bilingual Program
N= 53 ANOVA
Variables M SD M SD F p
Happy 47.69 9.90 43.53 12.76 4.26 0.0419
Sad 48.07 16.97 50.84 12.11 0.00 0.9684
Angry 51.02 16.34 49.84 12.70 0.54 0.4626
Total 48.93 11.89 48.06 9.70 1.24 0.2691
Interpretation
The graphic, visual and verbal representations of children’s relationships of self and significant
others used for this study provided evidence of their cognitive levels. The significant differences in
the performance of all students across the three modes of representation point to the critical role
that visual assessments play in providing essential and complementary information about children’s
cognitive functioning. Children’s representations of Self in drawing captured the three categories
proposed by James (1983): the constituents of self, the feelings and emotions and the actions.
Symbols used by students to portray interactions in their drawings provided concrete quantifiable
evidence for analysis and therefore measure of cognition.
Results indicate that children’s thinking patterns evolve in stages as Piaget (1936) uncovered or in
levels as Fischer (1980) delineated. Children’s competence in mental operations, which are
reflected in actions and symbols, provide indicators for identifying their level of performance.
Results also showed evidence of the influence of socio-emotional development in children’s mental
capacity as suggested by Gunnar and Sroufe (1991). Drawings provided representational (visual
and verbal) evidence of the various levels suggested by Fischer and Mascolo (1998) as students
included persons on their own, a person in a relationship with another, and a person as it functions
in society at large (Appendix R: 1-16).
1). How do the cognitive structures of monolingual and bilingual speakers compare when
performing in three distinct modes of representation: graphic composition, visual content, and
verbal narrative?
Results indicate that there is no statistical difference between the cognitive structures of
monolingual and bilingual speakers as measured by the levels of complexity achieved in three
modes of representation. This finding is consistent with Piaget’s claim that linguistic experiences
have little or no impact on a child’s psychological development. This finding is also in line with
Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001) who found that bilingual and
monolingual subjects develop at similar rates in biologically controlled mechanisms related to
linguistic acquisition.
I feel secure in saying that the cognitive development of monolingual and bilingual youngsters
evolves, at least, at similar rates. Even though, I noticed a hint that bilingual speakers tended to
score slightly higher than their monolingual counter parts in most of the measures. A subsequent
study, with larger number of participants, would be necessary to confirm this trend.
In general, all students scored higher levels of complexity in graphic composition than in visual
content, though scores in both modalities were highly correlated. This slight difference may be due
to the fact that the graphic scale is based on elements of composition (such as horizon lines and
perspective) and it was a four-point scale (though the analysis counted for that difference) and the
visual scale was based on the nature of relationships represented in the content of the drawings.
The strong correlation between graphic and visual scores indicates that students perform similarly
when being assessed in the visual domain and points to the validity of the Varon scale.
The five-point visual and verbal scales used in this study are closely inter-related. Scores for both
scales are derived from content analysis, which assesses portrayed relationships based on criteria
designed by Fischer (1980) in his skill theory. Both visual and verbal scores show weak correlation,
which indicates that students’ performances in different modes/domains fluctuate.
My anticipation is that the Varon visual scale used in this study will improve the “traditional” tool
used to measure artistic development (i.e. Lowenfeld’s scale based on graphic composition derived
from Piaget stages in the 50’s) in the same way that Fischer revised Piaget’s measures of cognitive
development with his proposed levels. The Varon scale is a timely response to Feldman’s
suggestion that stages of artistic development be reformulated based on current theories of
cognitive development.
When examining the differences across four groups, the difference between the English
monolinguals in the monolingual program and the bilinguals in the bilingual program is 0.28 points
in favor of the bilinguals. The difference between the English monolinguals and the bilinguals in the
monolingual program is 0.14 points. Spanish monolingual students in the bilingual program, on the
other hand, score consistently lower. The gap between bilinguals and Spanish monolinguals in the
bilingual program is 6.17 points when looking at all three representational modes. This group of
Spanish monolinguals also scored much lower in their Raven Progressive Matrices. The gap
between them and the English monolinguals in monolingual program is 3.65. The gap between
them and the bilinguals in the bilingual program is 3.70.
This gap between the Spanish monolingual speakers in the bilingual program and the rest of the
groups is most likely the effect of their recent immigration. These students have not had the same
enriched environmental and educational exposure as the other students. The drastically changed
environment and temporary unstable living conditions may contribute to a psychological block of
expression. In time however, the bilingual educational program will prove beneficial to them, as it
has been for other immigrants who become proficient bilinguals in the two-way bilingual immersion
program.
Students in both educational programs scored virtually the same. When examining all three
representational modes, the difference is 0.81 in favor of the monolingual program, especially in the
graphic composition with a 4.66-point difference. This gap is evidently produced by the lower
performance of the Spanish monolinguals students in the two-way bilingual program who happen
to be new immigrants with lower Raven scores.
2) How do the cognitive structures of monolingual and bilingual speakers compare when examining
their representations of three distinct emotions (happy, sad and angry)?
This dissertation found that there are no statistical significant differences across monolingual and
bilingual speakers when representing any of the targeted emotions. This is valid also when
comparing across the four groups. Nevertheless, there is evidence of a statistically significant
difference when representing happy (p=0.0419) when comparing students in the monolingual vs.
bilingual program.
Bilingual speakers score higher than monolingual students when representing a sad or an angry
event. The difference is higher when representing sad (4.36 points) than when representing angry
(0.82 points).
When representing a happy event in all three modalities, monolingual and bilingual speakers scored
virtually the same (0.18 points difference in favor of monolinguals). Angry events produce the
highest level of complexity in verbal narratives for both monolingual and bilingual speakers.
In examining the representation of emotions across four groups, “sad” produces the highest score
by the bilingual speakers in the two-way bilingual program. Next is “angry” which produces similar
scores by English monolinguals in the monolingual program and bilinguals in the two-way bilingual
program with a difference of only 0.59. Spanish monolingual students in the bilingual program
score consistently lower when representing all three emotions, with “happy” scoring the lowest.
When looking at the representation of emotions across two educational programs, the difference is
0.87 also in favor of the monolingual program. This difference is consistent and being affected by
the lower performance of Spanish monolingual students in the bilingual program.
3). How do monolingual and bilingual speakers compare in degrees of cognitive complexity as
manifested in the graphic composition of their drawings?
There is no evident statistical significant difference in the artistic development of monolingual and
bilingual speakers as measured by the analysis of their graphic compositions. Bilingual speakers
scored slightly higher (1.28 points higher) than monolingual speakers. There are also no statistical
significant differences found across the graphic representations of the four groups. Nevertheless,
there is evidence of a statistically significant difference (p=0.0379) when examining students’
graphic composition and comparing students in the monolingual vs. bilingual programs. This is due
to the lower performance of the Spanish monolingual students in the bilingual program, who, as
mentioned before, come to this country with a disadvantage. This finding supports evidence of
Feldman’s theory of universal vs. non-universal development where he states that there is no
guarantee that a child will move from stage to stage at the same rate as others if he or she has not
had similar environmental and educational enrichment. Since non-universal development is highly
dependent on environment and individual resources, immigrant children tend to lag behind in
development for as long as five or six years. This is line with the time that Cummins (1981), Collier
(1987) and Hakuta (1987) prescribe necessary for children to acquire full academic proficiency in
two languages.
4). How do they compare in the degrees of cognitive complexity manifested in the visual
representation of interaction in their drawings?
There is no evident statistical significant difference between the cognitive structures of monolingual
and bilingual speakers when examining any of the comparisons of students’ visual representations.
Even if the difference is not statistically significant, when the group is broken down into four
however, the monolingual speakers in the bilingual program scored the lowest with an 8.21-point
difference when compared with the bilingual speakers in the same program, and with a 4.74-point
difference with the monolingual speakers in the monolingual program.
Additionally, the fact that bilingual speakers scored (4.43 points) higher than the monolingual
speakers using visual assessment and (1.96 lower) than when using verbal measures of assessing
(see below) indicates how crucial it is to diversify the domains in which we assess. Visual measures
provide valuable and complimentary information.
When examining the scores in the visual representations across four groups, bilingual students in
the bilingual program and bilingual students in the monolingual program score similarly with 0.81-
point difference in favor of bilinguals in the bilingual program. The widest gap is between the
bilinguals and Spanish monolinguals in the two-way bilingual program again, with an 8.21-point
difference in favor of bilinguals in the bilingual program.
5). How do they compare in the degrees of cognitive complexity manifested through the verbal
descriptions of depicted interaction as measured by skills theory?
Findings indicate that there is no significant statistical difference between monolingual and bilingual
speakers’ cognitive development based on verbal measures. Monolingual speakers scored only
0.72 points higher than bilingual speakers. .
In examining data derived from students’ verbal narratives across four groups, I noticed a 5.11-
point difference between the English monolingual and the bilingual students in the monolingual
program. This finding supports Lambert’s (1987) finding that subtracting bilingual speakers of
their native language in school will negatively affects their cognitive development. Conversely, the
evidence points to the fact that schooling bilingual speakers in two languages will promote their
verbal development in the least.
When comparing both educational programs in students’ visual and verbal scores, students in the
bilingual program have the advantage. In the visual representations students in the bilingual
program are ahead by 1.47 points, and in their verbal narratives they are ahead by 1.74 points.
Conclusions
Since Fischer’s (1980) claim that development is quantifiable even if multifaceted and divergent
according to domain and individual differences, skill theory provided a suitable developmental
system (scale used to measure verbal narratives) for studying monolingual and bilingual children’s
cognitive profiles as measured by the narratives that described their drawings. The visual scale
used in this study was based on a system of role intersections and relationships derived from
Fischer (1980) levels of cognitive development. Children’s levels of performance started with
simple concrete structures and moved towards greater degrees of complexity that included
symbols and abstractions.
The scale used to measure graphic composition in this study was based on Lowenfeld’s stages of
artistic development. Due to the differences between the artistic (graphic) and cognitive conditions
(visual content and verbal narratives), I predicted greater advantage in visual and verbal
representations, rather than in graphic composition. However, results suggest that there are
significant differences in students’ performances in the three different modes, with evidence of
strong correlation between the graphic and visual scores. This leads me to conclude that the
Varon scale offers a reliable alternative, and possibly an improvement, for measuring cognitive
development through drawing.
My objective in this dissertation was to apply quantifiable measures used in the field of cognitive
psychology and to contribute to the field of artistic development by proposing an alternative scale
to Lowenfeld’s stages, which are used to measure graphic composition. The bilingual vs.
monolingual context provided me with an opportunity to further look into the impact of linguistic
development on children’s psychological and educational experiences.
Initially it was surprising to note the slight disparity between the representations of cognition in
psychology and in bilingual literature. In bilingual literature, cognition is closely tied to academic
performance, whereas in psychology, cognition is tied to a child’s biological maturation of ability to
perform a thinking task.
Bilingualism is developmental in that it takes 5 to 7 years for students (who do not grow up in
bilingual homes) to become bilingual. In this case, bilingualism fits into the “non-universal”
category proposed by Feldman (1985), and its development is dependant on environment and
educational experiences. When researchers on bilingualism such as Collier (1987, 1995 a- b),
Cummins (1981), Hakuta (1986, 1987,1990, 2001) Lambert, 1972), Nicoladis, E. Taylor, D.,
Lambert, W.E. & Cazabon, M. (1998); Thomas & Collier (2000-2002) write about the cognitive
advantages of bilingualism, they refer to context-specific school-related tasks not necessarily
associated with biological maturation. On the other hand, psychologists such as Piaget, in his
stages of cognitive development, refer to the universal biological maturation, which determines
children’s readiness to be able to perform a mental task.
Fischer combines the various points of view, integrating biological maturation, environmental
influences, and context specific experiences in a task to determine his levels of cognitive
performance. These are domain specific and conditioned by biological structures. However,
interactive experiences and probing from an expert will allow for optimal performance in a specific
task. “Skill theory clearly states that specific skills develop as a result of experience in particular
tasks (although there is a biological basis for the optimal level that a child can reach)”. (Fischer,
2004)
The findings of this study confirm that cognitive development precedes linguistic experiences and
that it is therefore universal, as Piaget had claimed. There is also evidence that the environmental
(cultural) influences have significant impact on cognitive development. Spanish monolingual
students who were recent immigrants to the United States often performed at lower levels of
complexity than other groups participating in the study.
Whereas bilingualism may not propel psychological or biological advantages, it also does not in any
way cause cognitive deficit, such as mental confusion. This finding is supported by studies
conducted by Petitto, Katerelos, Lecy, Gauna, Tetreault and Ferraro (2001).
Results suggest that there are no evident statistically significant advantages to bilingualism as
measured by all three representational modes. There is only a hint at actual cognitive
development advantages because bilingual speakers scored slightly higher than their monolingual
counterparts in most of the instances. In order to further explore this trend, a follow up study
with an increased number of participants would be necessary. In any case, this study revealed that
bilingualism does not in any way hinder individuals’ verbal or cognitive development.
In this dissertation, I demonstrated the developmental appropriateness of the Varon visual scales
in that they produced scores that corresponded with Fischer’s levels for the verbal narratives, and
where there was a significant statistical difference in the Raven Progressive Matrices, there was also
a statistically significant difference in the results produced by the visual scale.
In considering children’s differences in students’ skills performance associated with drawing, I
explored practical applications of drawing in measuring cognition. Drawings have not been used to
examine the differences between monolingual and bilingual speakers, except in my qualifying
paper. In that paper, I suggest that in the process of learning a second language, while not always
capable of vocalizing thoughts, students become more observant of visual elements such as
setting, interaction, and personal characteristics of people and environments (Varon, 2000). This
may have direct impact in the perceptual system (Eisner, 1967; Mehler, Pallier & Christophe, 1998;
Varon, 2000; Whorf, 1956).
Additionally, I explored how traditional measures of artistic development, such as the one created
by Lowenfeld, relate to measures of cognitive development. The fact that children’s scores in visual
content correlate with scores in graphic composition demonstrates that visual content as well as
graphic compositions successfully capture cognitive levels of performance in the same quantitative
way as do verbal narratives.
The implications of this study could impact the assessment processes used in schools today. This
study may be relevant to teachers who want to better understand their students as well as the
benefits and constraints of being monolingual or bilingual. Drawing is a natural activity for children
providing a comfortable assessment tool. The task and framework for analysis used in this study is
one that can be easily implemented in the classroom on a regular basis; it would be an alternative
way of evaluating children who may not perform as well in verbal or written assessments.
Due to the fact that I did not collect academic scores of students participating in this study, I am
not able to differentiate between academic and psychological performances of bilingual and
monolingual speakers as other researchers have done (Cazabon & Lambert, 1994; Cazabon,
Nicoladis & Lambert, 1998; Thomas &Collier, 1997, 2002). However, I am able to conclude that it
is crucial to diversify the domains in which we assess children. Using visual assessment in addition to
verbal measures provide valuable and complimentary information about children’s cognitive
structures.
Suggestions for future studies include furthering the application of the visual representational scale
to conduct a longitudinal developmental study that would include the drawings and academic
assessments of monolingual and bilingual speakers in elementary and middle schools.
