National Council of Teachers of Mathematics

Partitioning: The Emergence of Rational Number Ideas in Young Children

Author(s): Yvonne Pothier and Daiyo Sawada
Source: Journal for Research in Mathematics Education, Vol. 14, No. 5 (Nov., 1983), pp. 307-317
Published by: National Council of Teachers of Mathematics
Stable URL: http://www.jstor.org/stable/748675
Accessed: 15-10-2015 17:24 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/748675?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

National Council of Teachers of Mathematics is collaborating with JSTOR to digitize, preserve and extend access to Journal
for Research in Mathematics Education.

http://www.jstor.org

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

307

Journal for Research in Mathematics Education

1983, Vol. 14, No. 4, 307-317

PARTITIONING:THE EMERGENCE OF
RATIONAL NUMBER IDEAS
IN YOUNG CHILDREN
YVONNE POTHIER, Mount Saint Vincent University
DAIYO SAWADA, University of Alberta
The study sought to trace the emergence and differentiation of partitioning as a process
that leads young children to construct ideas of rational numbers. The method involved
clinical sessions with children as they manipulated materials while solving tasks designed
to reveal partitioning processes. The data analysis led to a theory concerning the
emergence of partitioning capability, and conceptual structures that undergird the capability were identified. A five-level theory of the development of the partitioning process is
presented.

Key position papers by Kieren (1976, 1980) have created a new theoretical
context for inquiring into the child's acquisition of rational number concepts.
Recent work within the new perspective (Kieren & Nelson, 1978; Kieren &
Southwell, 1979) has lent support to Kieren's position.
Basic in Kieren's perspective is the process of dividing a whole into parts, a
process variously referred to as partitioning, subdivision, or dissection. The
purpose of the present study was to trace the emergence and differentiation of
the process of partitioning as revealed in children's attempts to subdivide a
continuous whole into equal parts.
METHOD
The method involved a clinical interaction with a child as he or she
grappled with tasks specifically designed to reveal partitioning capabilities.
As the child grappled with the tasks, the researchers grappled with the even
more complex task of attempting to make sense of the child's behavior. We
hoped to discover underlying patterns that would help us trace the emergence
and differentiation of the partitioning process.
The method can be characterized as a clinical interaction technique set
within a discovery paradigm. The interactive encounters were designed to
yield insights rather than to verify hypotheses; validity, as opposed to
generalizability, was our concern.
The interaction was characterized by flexibility in questioning. The initial
question for each task was standard, but the subsequent questions, although
following a general pattern, were varied, as were the numbers in the problem,
depending on the behavior of the child. The method permitted probing,
This report is based on the first author's doctoral dissertation, completed at the
University of Alberta in 1981 under the supervision of the second author.

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

308

Partitioningand the Emergenceof RationalNumber

following through a child's peculiar behavior, expanding on a question, or

altering the problem.
Sample
The 43 children who formed the sample were from kindergarten and the
first three grades in a primary school in Alberta. The selection of respondents
followed the method of theoretical sampling (Glaser & Strauss, 1967) that
requires that no a priori decision be made regarding sample size; rather, the
selection continues throughout the period of data collection. The children, all
acquainted with the interviewer, were chosen on the basis of the interviewer's
belief that they would respond freely to the task during a clinical session. The
sampling continued until sufficient key insights had emerged and a satisfactory degree of conceptual integrity connected the various insights. At the close
of the clinical sessions, the sample included 8 kindergarten, 8 Grade 1, 12
Grade 2, and 15 Grade 3 children. Their ages ranged from 4 years, 11 months
to 9 years, 8 months.
Task
Although five partitioning tasks were devised for the study, one task, the
Cake Problem, proved to be highly effective both in enabling children to
demonstrate their partitioning capabilities and techniques and in giving us
key insights into the partitioning process. (For details of the other tasks, see
Pothier, 1981.)
The materials for the Cake Problem consisted of one circular and four
rectangular "cakes" of various dimensions, a circular "cookie," a supply of
sticks, and a set of miniature dolls. The rectangular cakes were white
Styrofoam models, 5 cm in height and topped with a piece of plush gold rug
glued to the surface. The dimensions of the rectangular cakes were 10 cm by
25 cm, 15 cm by 20 cm, 20 cm by 25 cm, and 10 cm by 15 cm. The circular
cake, 20 cm in diameter and 2.4 cm in height, was constructed of the same
materials as the rectangular cakes. A circular piece of rug, 16 cm in diameter,
was also used and was referred to as a "giant cookie." The sticks, which
served to demonstrate cuts on the cakes and the cookie, were white plastic
molding, 1 mm thick, cut in strips 0.5 cm wide and of lengths 10, 19, and 27
cm. The complete supply of sticks was available for the child's use when
partitioning each cake or cookie for any number of dolls among whom it was
to be shared.
Procedure
The setting for the Cake Problem was a birthday party, and the child was
asked to demonstrate how he or she would cut the cake so that each person at
the party had the same amount. The question posed to the child was, "How
would you cut the cake so that each one has as much?" After one way had
been demonstrated, the child was requested to show another way to cut the
cake so that each person had the same amount.

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

YvonnePothierand Daiyo Sawada

309

The child was first requested to partition the cake for 2 people. In most
cases, the subsequent partitions were for 4, 3, and 5 people. The order of
presentation varied, however, and not everyone was asked to partition the
cakes for the same numbers of people. The older children were asked to make
more partitions than the younger ones. Additional sequences for the older
children were 4, 8, and 16 people; 5, 10, and 20 people; 3, 6, and 12 people.
The circular region was partitioned first; that is, the children were presented first with the giant cookie and subsequently with the rectangular
cakes. The round cake was generally employed for further partitioning of a
circular region following work with the rectangular cakes.
In addition to video and audio recording, any distinctive behavior observed
or impressions made were recorded in written form during the sessions.
Analysis Scheme
The analysis was based on a method suggested by Glaser & Strauss (1967):
Theoretical sampling is the process of data collection for generating theory whereby the analyst
jointly collects, codes, and analyzes his data and decides what data to collect next and where to
find them, in order to develop his theory as it emerges. This process of data collection is
controlled by the emerging theory, whether substantive or formal. (p. 45)

Accordingly, a three-stage analysis scheme was used.

Immersion stage. The first stage was a continuous process of analysis
characterized by an intense focus on the child's behavior. Its ongoing and
flexible nature provided the direction for immediate probing and for variations in the task.
Reflection stage. During the period of data collection, time was daily
devoted to reflecting on the day's interactions and viewing, transcribing, and
analyzing taped sessions. The interviewer kept a journal of insights gained
and inferences made. The reflective process provided direction for further
data collection.
Documentation stage. The third stage of the analysis was conducted after
the data had been collected and comprised a systematic examination of the
tapes and other records made during the first two stages. Charts, lists, and
tables (some presented below) were constructed for the purpose of verifying
the interpretations and for gaining further insights.
FINDINGS
The theoretical account given below is a formalization of the insights that
emerged over the three stages of the data analysis. Realizing that the formalization is a rather severe distillation of the data, we first identify some
constructs that play a key role in the theory, and we follow the account of the
theory with excerpts from the transcripts.
To trace children's partitioning capabilities as precursors of rational
number acquisition, the relevant mathematical constructs, contrary to expectation, are not those of rational number. Rather, we propose that the key

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

310

Partitioningand the Emergenceof RationalNumber

theoretical constructs come primarily from number theory and secondarily

from geometry. Our description of the emergence and differentiation of the
partitioning process hinged critically on the use of bipolar number-theoretic
concepts: odd/even, prime/composite, factor/multiple. We synthesized the
bipolar constructs into a final level of partitioning capability suggested by the
fundamental theorem of arithmetic. We made refinements in each level of
capability by using concepts from motion geometry: translation, reflection,
rotation, similarity, congruence, and symmetry. The use of theoretical constructs derived from number theory and geometry does not imply that
children are aware of these constructs. We simply find them valuable in
making sense of the clinical interactions we had with the children.
We propose a five-level theory that describes the development of the
partitioning process as we see it. Four levels are grounded in the data; the fifth
level follows logically, but remains hypothetical.
Four Levels of Partitioning Capabilities
The first four levels are outlined below in terms of three distinctive characteristics: (a) the construct, or key concept, developing during the level; (b) the
algorithm, or procedure, employed to produce the partitions; and (c) the
domain, or extent, of the partitioning capabilities within the level.
Level I: Sharing

"*Construct-breaking; sharing; halving

"*Algorithm-allocating pieces ("a piece for you")
"*Domain-social setting; counting numbers
Level II: Algorithmic halving
"*Construct-systematic partitioning in two
-no notion of equality
"*Algorithm--repeated dichotomies
"*Domain-one-half and other unit fractions whose denominators are
powers of 2
Level III: Evenness
* Construct-equality; congruence
-repeated dichotomies becoming meaningful
* Algorithm--halving algorithm; geometric transformations
-extension of algorithmic halving to doubling the number of
partitions and adding two parts
* Domain-unit fractions with even denominators
Level IV: Oddness
* Construct-even and odd
-search for a new first move
-use of the new first move
-geometric transformations

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

YvonnePothierand Daiyo Sawada

311

"*Algorithm--exploratory measures; trial and error

-counting; one-by-one procedure
unit fractions

"*Domain-all

Interpreting the Levels

Level I: Sharing. The action of partitioning an object or a set of objects is
learned by a child in a social setting (Kieren, 1980). At an early age, the child
witnesses situations in which things are shared and hears expressions such as
"Break (or cut) it in half" and "Here's half." Later, the child participates in
sharing activities, learns how to share, and eventually uses language to
describe the sharing action and the result of sharing. But as far as rational
number is concerned, this is rote learning. The child does not know the
meaning of half in a number sense, and the necessary characteristics of
evenness and one of two parts are not understood. This is shown by the child's
use of the term half in expressions like break it in half in four pieces and
split it in half in three pieces. Also, some children use the term half exclusively
as an operation. An example is Lon, aged 5 years and 6 months (5; 6), who
stated that the cookie was broken in half but called each piece a broken
cookie. To the query "What could you call that piece?" he responded, "I
don't really know."
At the Sharing level, the child begins to learn a halving mechanism, that is,
how to construct a line through the middle of a region. Using the mechanism,
the child can usually partition rectangular and circular regions in halves and
fourths. However, some attempts result in uneven shares, in more parts than
required, or in the region not being used up. Samples of these partitioning
attempts are depicted in Figure 1. At this level, partitioning continuous
quantities simply means allocating "pieces," and an equal number, regardless
of size, is considered a "fair share."
Level II: Algorithmic halving. A second level is marked by the mastery of
the doubling process that enables the child to partition rectangular and
circular regions not only in halves and fourths but also in eighths and
sixteenths. In other words, the child can double the number of parts to obtain
fractional parts whose denominators are powers of 2. The child uses the
acquired tool in an algorithmic manner, but there is no concern for equality.
Successive partitions are performed systematically so that each part in turn is
bisected, thus doubling the previous number of parts. Because the algorithm
"works" so well, it is used without careful scrutiny of the partitions it
produces. The procedure is often used in situations where it may be inadequate or unsatisfactory. For example, a diagonal line can be used to
partition a rectangular region evenly in two, so the child either performs
repeated "dichotomies" in an effort to obtain fourths (see Figure2a) or thirds
(see Figure 2b) or mixes two halving lines (see Figure 2c). Vertical parallel
lines that work in a rectangular region are used in a circular region to produce
thirds, fourths, or fifths (see Figure 2d). In such instances, the parts are

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

312

Partitioning and the Emergence of Rational Number

Figure 1. Level I partitioning behavior.

(a)

(c)

(b)

(d)

Figure 2. Level II partitioning behavior: Thirds, fourths, and fifths.

declared to be fair shares. Examples of attempts at eighths and sixteenths are

portrayed in Figure 3.
Level III: Evenness. A third level of achievement is attained when the child

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

Yvonne Pothier and Daiyo Sawada

313

Eighths

Sixteenths

Figure 3. Level II partitioning behavior: Eighths and sixteenths.

attends to the size and shape of the parts and critically examines them to
determine if they are equal, or even. At this level, partitions are correctly
classified as "fair" or "not fair," although some children may not be able to
produce equal shares (e.g., thirds or fifths).
Level III is a major breakthrough in the child's thought. Algorithmic
halving becomes more meaningful; equality (and evenness) is now a critical
characteristic in designating fair shares. Diligent attempts are made to construct equal pieces. For many children partitions that do not look the same are
designated "not fair." Moreover, the equality concept, together with geometric motions, enables the child to produce partitions for numbers other than
powers of 2. Even numbers of parts can now be obtained. For example, to
obtain sixths, the child typically partitions a circular region into fourths,
bisects two sections, and then rotates two sticks to obtain six equal parts (see
Figure 4a). In a rectangular region, sixths can be attained from fourths by
translating a halving stick and placing an additional stick parallel to it (see
Figure 4b). In a similar manner, tenths can be obtained from eighths, and
twelfths from sixths.
(a)
2
1

3.

"1

(b)

Figure 4. Level III partitioning behavior: Sixths.

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

314

Partitioningand the Emergenceof RationalNumber

This primitive use of geometric motions enhances algorithmic halving so

that the child can obtain unit fractional parts whose denominators are even
numbers. We refer to this more sophisticated capability as the halving algorithm. Although we did not test this conclusion, we consider the halving
algorithm to be composed of two things: the doubling of a number of parts
and the attainment of additional parts by adding another stick and making
rotational moves.
In summary, the Evenness level is marked by two notions of evenness: (a)
the equality, or evenness, of parts, and (b) the ability to partition unit
fractions with even denominators. Children at this level are unable to obtain
thirds and fifths. Typical attempts are depicted in Figure 5.

Figure 5. Level III partitioning behavior: Thirds and fifths.

Level IV: Oddness. Following the Evenness level, there is a period in which
the child becomes increasingly aware of the inadequacy of the halving algorithm. The child learns that by beginning with a halving line, some fractional parts cannot be attained efficiently in either a rectangular or a circular
region. The problematic numbers are three and five, or odd numbers as
shown earlier in Figure 5. Some children label these as not even or the hard
ones. When asked to make an odd number of parts, some children respond

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

YvonnePothierand Daiyo Sawada

315

with statements such as "I can't get it because it's not even" or "It can't be
done."
Hof (8; 10) and Bro (8; 5) are children who had reached Level IV.
When partitioning a rectangular cake for seven people, Hof stated, "This
one's an odd number.... I can do even numbers, but it's hard to do odds." He
defined even and odd numbers: "It means like, if you're countin' by twos,
they're even numbers. If you count by ones, like, I should say by threes-one,
three, like that, they're odd numbers."
While partitioning a round cake, Bro stated that three and five parts were
"hard to make." Asked, "What's hard about three and five?" she replied,
"Well, if you try to make something out of 'em, you already made them." She
explained further: "Well, you try to make five, but then you get six." In
attempting thirds, she stated, "Like if you make three like this [a horizontal
diameter cut and two vertical radius cuts], you'd get four, so, like if you go
like this [moving the sticks], you still get four. If you go like this [one
horizontal diameter cut], it's two halves."
Realizing that a different initial cut is necessary to partition into thirds and
fifths, the child actively searches for a different move, thus freeing the partitioning from the dominance of the halving algorithm. For a circular shape, the
first move is a cut along a radius of the circle, whereas for a rectangular shape,
it is a cut in a vertical or horizontal line in a position other than the middle.
The child uses this new move to obtain fractional parts whose denominators
are odd.
In the case of an odd number of parts, the child uses a counting algorithm
to guide the partitioning; that is, pieces are produced one by one, and
counting is used to keep track of results. Such a strategy requires frequent
readjustment of the partitioning lines to obtain equal shares. Having attained
the oddness level, the child is able to partition rectangular and circular regions
into any number of parts depending on his or her counting ability.
Level V: Composition. The counting algorithm used in Level IV is suitable
for small odd prime numbers. However, it is an awkward algorithm for large
numbers such as 9 and 15 because equal parts are not easily attained. In time,
the child will probably question the one-by-one procedure and look for a
different method. For composite numbers such as 9 and 15, there is clearly a
more efficient strategy.
To obtain ninths, a region can first be partitioned into thirds and then each
part trisected. In like manner, fifteenths can be composed by trisecting fifths.
The possibility of alternative partitioning sequences demonstrates that the
choice of a factor for a first move does not affect the result: an illustration of
the fundamental theorem of arithmetic. A child who proceeds to partition
composite numbers in this manner employs a multiplicative algorithm. Such
a child has reached the Composition level and can now efficiently construct
all unit fractionst
As stated earlier, the fifth level is not based on data; none of the children in

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

316

Partitioningand the Emergenceof RationalNumber

the study used a multiplicative algorithm when partitioning the regions. This
level is strictly an implication of the theory-a level that would emerge, we
hypothesize, with older children.
DISCUSSION
We propose that the acquisition of the partitioning process is the culmination of a gradual progression through five levels. Each level is distinguished by
certain conceptual characteristics, procedural behaviors, and partitioning
capabilities. In particular, the partitioning process is mastered through the
successive attainment of five subsets of the unit fractions: the fraction 1/2,
fractions whose denominators are powers of 2, fractions with even denominators, fractions with odd denominators, and fractions with composite
denominators.
A child first learns to partition in two; then, with the acquisition and
eventual mastery of the halving algorithm, in powers of 2; then, with the use
of geometric motions, in even numbers. Partitioning in odd numbers follows
the learning of a first move other than a median cut. With the discovery of the
new first move, children are able to partition in thirds, fifths, and other odd
numbers; thus, thirds and fifths are achieved together. The algorithm involves
counting, and equality of parts is usually achieved by rotational (for circular
shapes) and translational (for rectangular shapes) moves.
For continued mastery of partitioning, ideas of odd and even appear to play
a role. Children who constructed thirds or fifths but not both appeared not to
be cognizant of the oddness quality. Their use of a nonhalving first move
seemed accidental.
In attempts to partition rectangular and circular regions into thirds and
fifths, the dominance of the halving algorithm was evidenced. Many children
seemed incapable of deviating from employing a halving line as the initial
"cut." A few who began with a line other than a median line did so only after
several trials or with guidance. Since young children ordinarily master halves
and fourths prior to thirds (Hiebert & Tonnessen, 1978; Piaget, Inhelder, &
Szeminska, 1948/1960), the children in the sample were first asked to partition the regions into an even number of equal-sized parts. Would the results
have been different if partitioning into an odd number of parts had been the
first activity? Would the halving algorithm have been so dominant, or is the
behavior a sequence effect? This issue is the focus of our current research.
CONCLUSION
The goal of the study was to gain an understanding of the partitioning
process as it relates to the development of rational number ideas. From an
interpretation of the partitioning behavior of young children, a theory has
been generated wherein foundational constructs are identified and levels
characterizing distinct partitioning capabilities are delineated. Within each
level, developing concepts and algorithms, together with increasing domains

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions

YvonnePothierand Daiyo Sawada

317

of partitioning capabilities, are described. Our results, like those of Hiebert

and Tonnessen (1978), are in substantial agreement with the results of Piaget,
Inhelder, and Szeminska (1948/1960). Our main departure has been in
discovering the appropriateness of number-theoretic concepts in making
sense of the emergence and differentiation of the partitioning process.
REFERENCES
Glaser, B. G., & Strauss, A. L. The discovery of grounded theory: Strategies for qualitative
research. Chicago: Aldine, 1967.
Hiebert, J., & Tonnessen, L. H. Development of the fraction concept in two physical contexts:
An exploratory investigation. Journal for Research in Mathematics Education, 1978, 9,
374-378.
Kieren, T. E. On the mathematical, cognitive, and instructional foundations of rational numbers.
In R. A. Lesh & D. A. Bradbard (Eds.), Number and measurement: Papers from a research
workshop. Columbus, OH: ERIC Information Analysis Center for Science, Mathematics and
Environmental Education, 1976.
Kieren, T. E. The rational number construct-Its elements and mechanisms. In T. E. Kieren
(Ed.), Recent research on number learning. Columbus, OH: ERIC Clearinghouse for Science,
Mathematics and Environmental Education, 1980.
Kieren, T., & Nelson, D. The operator construct of rational numbers in childhood and adolescence-An exploratory study. Alberta Journal of Educational Research, 1978, 24 (1), 22-30.
Kieren, T., & Southwell, B. The development in children and adolescents of the construct of
rational numbers as operators. Alberta Journal of Educational Research, 1979, 25 (6), 234247.
Piaget, J., Inhelder, B., & Szeminska, A. The child's conception of geometry (E. A. Lunzer,
Trans.). New York: Basic Books, 1960. (Original French edition, 1948.)
Pothier, Y. Partitioning: Construction of rational number in young children. Unpublished
doctoral dissertation, University of Alberta, 1981.

[Received May 1982; revised January 1983; revised June 1983]

This content downloaded from 200.37.4.202 on Thu, 15 Oct 2015 17:24:18 UTC
All use subject to JSTOR Terms and Conditions