You are on page 1of 29

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222050266

Writing in calculus and reflective abstraction

Article  in  The Journal of Mathematical Behavior · December 2002


DOI: 10.1016/S0732-3123(02)00129-3

CITATIONS READS
20 438

1 author:

Laurel Cooley
City University of New York - Brooklyn College
16 PUBLICATIONS   265 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Laurel Cooley on 09 January 2019.

The user has requested enhancement of the downloaded file.


Journal of Mathematical Behavior
21 (2002) 255–282

Writing in calculus and reflective abstraction


Laurel Cooley∗
Department of Mathematics and Computer Studies, York College, City University of New York, Jamaica, NY 11451, USA

Abstract
Reflective abstraction is a term that has been used in various, yet similar, ways in mathematics education. The use
of this term is examined with a final definition honed from this group and used for the purposes of this study. For this
study, reflective abstraction was defined as a mechanism for the isolation of particular attributes of a mathematical
structure that allows the subject to construct or reconstruct knowledge that is new; that is, not previously known.
With this definition of reflective abstraction as a basis, seven formal writing assignments were developed for a
calculus I class. Students completed these writing assignments and returned them in a journal which was read and
responded to by the professor. In some cases, students were required to rewrite their answers.
This group of students was at a public 4-year college in a metropolitan area. The majority of the students
spoke English as a second language. The college had recently begun implementing a writing across the curriculum
component and this study was born out of these efforts.
Since none of the students had studied calculus prior to this class, any construction of knowledge about the
calculus concepts was assumed to have been developed during the course of this class. The writing from the
students demonstrated that they had reflected on the calculus concepts and constructed meaningful ideas about
functions and calculus. This understanding was sometimes shown through personal analogies and sometimes in
a more formal way. The students were actively engaged in writing and produced thoughtful descriptions in most
cases. The process proved to be an important conduit for information between professor and student as well as a
strong tool in dissuading misconceptions and promoting reflection in their thinking.
© 2002 Elsevier Science Inc. All rights reserved.
Keywords: Writing and mathematics; Language and mathematics; Cognitive theory; Conceptual knowledge; Calculus;
Constructivism; Epistemology

1. Introduction

This study involved designing and integrating writing assignments into a calculus class with the goal
of promoting reflective abstraction. For some time, researchers have had a strong interest in the ways that

Tel.: +1-718-262-2545.
E-mail address: cooley@york.cuny.edu (L. Cooley).

0732-3123/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 7 3 2 - 3 1 2 3 ( 0 2 ) 0 0 1 2 9 - 3
256 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

language and mathematics learning interact. There has also been a growing awareness of the importance
of writing and the development of mathematical concepts. Two key questions guided this study: in
mathematics, can writing be helpful in the learning process? More specifically, can writing assignments
assist students in the process of reflective abstraction?
The objective of this study was to determine if writing assignments in a calculus I class would have any
effect in promoting reflective abstraction. Students were directed to complete writing assignments that
were designed to focus the student on particular topics. These writing assignments were then collected
and evaluated by the professor who would return them with comments, sometimes asking the students
to rewrite them. At the end of the semester, the journals in which the students kept all of their written
work were collected. These journals were evaluated for evidence of whether the process of reflective
abstraction had taken place. Examples of these instances are given for each of the seven formal writing
assignments. In addition, because the writing assignments facilitated communication between student
and professor, it became clear that this was a valuable tool for determining when students were having
misunderstandings. Examples of these types of written responses are given as well.

2. Relevant literature

There are two main aspects to this study. The first concerns reflective abstraction and the other is writing
in mathematics. Therefore, the literature focuses on these two themes, or closely related issues.

2.1. Reflective abstraction and construction of knowledge

In order to discuss abstract thought and reflective abstraction, these terms first need to be defined. Many
papers have been written on these subjects, or subjects closely related to them, with just as many varied
descriptions. Here, I will concentrate on constructivist writings, and the interpretations of these concepts
from this theoretical viewpoint. Then, I will offer my own synthesis which I used to inform the design of
the writing assignments.
An originator of constructivist ideas, Beth and Piaget (1966) spoke mainly of two forms of abstraction,
empirical and reflective. Empirical abstraction, the less sophisticated of the two, consists of eliciting
common attributes from a category of objects. Empirical abstraction refers to objects that are obvious
to the subject, who records certain properties in order to excerpt and analyze them, abstracting from
the perceived objects. However, Piaget points out, in physics and a fortiori in mathematics, there is a
second form of abstraction, called “reflective abstraction.” This second type of abstraction applies to
the subject’s actions and operations as well as to the schemata which it guides the subject to construct.
Reflective abstraction may be understood in two different, but associated, ways. To begin with, it may
be understood as “reflection” from a lower to a higher level (for example action to representation), or
as a reconstruction, on a new level, of what is elicited from the preliminary one. Therefore, reflective
abstraction — abstraction starting from actions and operations — differs from empirical abstraction in
that reflective abstraction is fundamentally constructive. (Beth & Piaget, 1966, pp. 188–189).
When properly understood, reflective abstraction is the mechanism which spurs development of in-
tellectual thought. Reflective abstraction differs from empirical abstraction in that it deals with actions
as opposed to objects, so that it is concerned, not so much with the actions themselves, but with the
interrelationships among the actions, which Piaget (1976, p. 300) called “general coordinations.”
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 257

The Piagetian process of reflective abstraction is thus understood as the following: the subject observes
the results of actions performed upon any objects resulting in logico-mathematical experience. The conse-
quence of this experience is interpreted by the schemas of the actions developed by the subject. In order to
observe this output, the subject carries out other actions, using the same schemas as those the interaction
of which must be considered. However, the form is new for the subject because the logico-mathematical
experience teaches her or him something he or she was not conscious of formerly. Therefore, the
abstraction by means of which the subject brings forth new knowledge involves construction. This
construction, or reconstruction, replaces experience or empirical procedures for the subject at a new
plane.
Twenty years later, Dubinsky and Lewin (1986) expanded on Piaget’s work, especially his concept of
genetic epistemology. They explain that an educator who is attempting to teach new conceptual material,
as in mathematics, is in effect trying to induce cognitive development. They focus on Piaget’s ideas of
“equilibration” and “reflective abstraction.” Equilibration is defined as the process by which the subject
tries to understand a concept by putting it in the context of her or his overall cognitive system. This
attempt to cognitively construct an understanding is through the process of reflective abstraction, which
Piaget defined.
In another viewpoint of abstraction, Dienes’ (1971, 1978) describes it in terms of concentrating on
common properties. His psychological studies of children were strongly influenced by his perception of
the nature of mathematics which he believed was a discipline created through the centuries by means of
successive abstractions. Therefore, he suggests that abstraction is the main characteristic of mathematics,
as well as the key point in the learning of mathematics. He offers as examples some areas of mathematics,
such as number theory, which can be developed by an in-depth study of a specific situation such as the
natural numbers. He admits, however, that concentrating only on common properties can be limiting.
Mathematics can have important commonalities in some cases of a concept, but not all. Instances where
commonalities do not exist are also important for further understanding and development.
All of this points to an ethereal quality of reflective abstraction which makes it difficult to define
in concrete terms. Something that is not easily defined may not be easily recognized. The abstraction
of mathematics is one of the main obstacles to its appreciation. For example, Mason (1989) finds that
students of mathematics often say that they find it abstract, and hence dislike it. Yet it is the abstraction
of mathematics that gives so much pleasure to mathematicians. Along the same lines, Eco (1988), in
discussing the meaning of aesthetic in the work of St. Thomas Aquinas, writes: “Aesthetic seeing involves
grasping the form in the sensible. It therefore occurs prior to the act of abstraction, because in abstraction
the form is divorced from the sensible.” (p. 193)
Mason’s insight is useful because it helps us to understand a student’s experience of mathematics. The
student’s sense of abstraction is removed from or divorced from reality or meaning. In a related manner,
Borasi (1984) proposes that students’ difficulties with abstraction have little or nothing to do with the
process of abstraction. She explains, for example, that the primary objective during a history class was
not to have students make a rapid advancement through levels of progress. Borasi (1984) notes that even
though number theory can be treated abstractly today, this is, in part, because Cantor developed his set
theory more than 2000 years after mathematicians started to study and discover the properties of the
numbers 1, 2, 3, . . . . However, students are expected to progress rapidly through mathematics, learning
what others have done, and not necessarily participating in the abstraction process. Ultimately, Borasi
defines abstraction as: “a class of situations and each situation belongs to this class because of a certain
property it has and any other properties it might have are, for the moment, considered as irrelevant” (1984,
258 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

p. 14). In a like manner, Tall (1988) states: “abstraction (as) the isolation of specific attributes of a concept
so that they can be considered separately from the other attributes.”
Correspondingly, Mason (1989) proposes that the use of the word abstract in mathematics by both
students and professionals refers to a common experience, an extremely brief moment. There is a “delicate
shift of attention” from seeing an expression as an expression of generality, to seeing the expression as
an object or property. Thus, abstracting exists between the expression of generality and the manipulation
of that expression while, for example, constructing a convincing argument. He argues that when the shift
occurs, it is hardly noticeable. To a mathematician, this is very natural. Hence, he states, when the shift
does not happen for the student, it blocks advancement.
Dienes, Mason, and Borasi, all discuss abstraction in some way in terms of the common properties
of a group of objects. Recognizing common properties may lead to specific attributes or may assist in
constituting generalizations. For example, Thurston (1990) believes abstracting is intimately linked to
generalization. He describes the general nature of the results that can be obtained through abstraction
as one of its principal incentives. Another primary motive is the achievement of synthesis. However,
for Thurston, the difference between simple generalization and abstraction is that generalization usually
involves an expansion of the individual knowledge structure, adding more objects to the category, while
abstraction is likely to involve a mental reconstruction.
Thurston argues that abstraction thus contains the potential for both generalization and synthesis and
that these two processes define its purpose. The nature of the mental process of abstracting is, however,
very different from that of generalizing and synthesizing. Abstraction, as he explains it and as similarly
discussed by previously mentioned authors, is first and foremost a constructive process — the building
of mental structures from mathematical structures, i.e., from properties of and relationships between
mathematical objects. He also believes this process depends on the isolation of appropriate properties
and relationships. It requires the ability to shift attention from the objects themselves to the structure of
their properties and relationships.
Finally, similar to Mason’s shift of attention, Thurston explains that such constructive mental activity
on the part of a student is heavily dependent on the student’s attention being focused on those structures
which are to form part of the abstract concept and drawn away from those which are irrelevant in the
intended context. In other words, the structure becomes important while irrelevant details are being
omitted, thus reducing the complexity of the situation.
Southwell (1988) explains the processes involved in reflecting on experience as association, integra-
tion, validation, and appropriation. She, like all of these researchers, believes that new ideas need to be
associated with knowledge that has already been acquired. She explains that these associations need to be
integrated methodically into a new whole, and that the new ideas must be validated by the learner’s previ-
ously held ideas. Then for some, though not all, the integrated ideas become a part of their value system.
Dubinsky (1991a, 1991b) describes reflective abstraction as a concept introduced by Piaget to describe
an individual’s construction of logico-mathematical structures during the course of her or his cognitive
development. He notes two important observations made by Piaget: first, that reflective abstraction has
no absolute beginning, but rather is present at the very earliest stages in the coordination of sensorimotor
structures (Beth & Piaget, 1966, pp. 203–208); and secondly, that it continues through higher mathematics
to the extent that the development of mathematics from antiquity to the present day may be considered
as an example of the process of reflective abstraction (Piaget, 1985, pp. 149–150).
A synthesis of these ideas produced the definition of reflective abstraction that is used for this study:
reflective abstraction is a mechanism for the isolation of particular attributes of a mathematical structure
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 259

that allows the subject to construct or reconstruct knowledge that is new; that is, knowledge not previously
known. A feature of reflective abstraction is that it clarifies and organizes logico-mathematical experiences
in such a way as to recognize both nuances and broad generalizations among them. Any new constructions
will be associated with knowledge the subject already has. The subject orders or re-orders a class of
situations with the characteristics of the current object so that the new knowledge fits with previous
schemas, or the previous schema has been reconstructed. The new generalization occurs precisely because
of a mental construction or reconstruction.

2.2. Writing and other reflective practices in mathematics

There have been many articles published in recent years on the use of writing in mathematics. The
authors included here used writing in some type of reflective manner. Their results and advice were used
to develop the writing component of this study.
Mason (1989) studied students who were given problems to solve with instructions that directed their
focus to the problem and were instructed to write their answers to the problem. The pointers directed
the students towards the “shift,” as Mason refers to it, in order to focus on abstraction and reflect on a
property of the problem which could lead to abstraction. At some points, this may have meant focusing
on the particular and the detail. At another moment, this required letting go of specific characteristics, a
“drawing away” from specificity. The experienced mathematician does this without thinking, and often
leaves the student behind. The point was to try to get the students also to make this delicate shift of focus.
Mason points out that mathematicians, in hopes of helping students, often look for devices, or physical
objects (such as diagrams or images) in which they see their own manifestations of their mathematical
representations, their abstraction. They then offer their “representations” to students, perhaps forgetting
that students have their own perspectives, which may be different. Hence, the student would be better off
if they were helped to draw their attention to what is being stressed, in order to assist them in seeing a
generality, and then expressing that generality to make the abstract shift in which the generality becomes
an object. Once abstraction is recognized as a shift of attention, students may be helped by activities
which direct them to generalize and to express this generalization in their own terms. Then the teacher
can bring the student’s focus to the process in which one can draw explicit attention to the movement in
which the contents become objects.
Bishop (1985) discusses the idea of what he refers to as “constructive alternativism.” He believes that
the teacher is the most important factor in the whole educational enterprise and, therefore, that it is very
important to research the decisions made and activities planned by the teacher.
Bishop claims that focusing on teaching activities is a significant move away from the idea of “teaching
methods.” The idea of a teaching method creates a distinction between it and the mathematical content.
The notion of a mathematical activity relates to both topic and process, and is a unit of both method
and curriculum. Bishop focuses mainly on spatial activities, but states that this can be embedded in the
more general category of mathematical activities, which he believes is a very important area of research.
Piaget (1975, p. 16) also stated that the teachers are necessary to create situations and initial tools which
are useful problems to the child. In addition to this, he explained that the teacher should also provide
counter-examples that compel reflection and reconsideration of overhasty solutions.
Bishop emphasizes the personal nature of the meaning of any new mathematical idea. A new idea
is meaningful to the extent that it makes connections with the individual’s present knowledge. It may
connect with the individual’s knowledge of other topics and ideas in mathematics and may also connect
260 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

with knowledge of other subjects outside of mathematics. The ultimate problem for teaching mathematics
then is not that of rigor, but of the development of meaning, of the existence of mathematical objects.
The goal of teaching then is in sharing and developing mathematical meaning. Rather than spending a
majority of time thinking about content, knowledge, and mathematical topics, teachers should be thinking
about the student activities in class. A focus on these mathematical activities can improve the situation
and put the activity of the student at the center of the teacher’s concerns.
Bishop argues that communication is a key element in the mathematics classroom. He is not referring
to a new construct, but the type of communication where mathematical meaning is discussed. Meaning
and understanding are about the connections one has between ideas. Communication in the mathematics
classroom is, therefore, about sharing mathematical meanings and connections. We can only share ideas by
exposing them. Important vehicles for this are talk, symbolism, uses of diagrams, examples in different
contexts, analogies and metaphors, and written accounts and descriptions. Interactive communication
between the teacher and student is necessary to learn about the students’ analogies, contexts, examples,
etc. and enable those to be exposed and shared. The ideas of interactive communication will encourage
students to take more part in sharing of mathematical meaning.
Along these same lines, von Glasersfeld (1996) writes that in order to teach abstract ideas, the instructor
must generate experiential situations in which the students themselves can make the necessary abstrac-
tions. In order to encourage such abstractions, the instructor must be successful in establishing a common
language with the students, that is a language of carefully negotiated and coordinated meanings, or as he
calls it, a consensual domain. He also states that, from the constructivist view, it is not helpful to assume
the students’ answers are simply wrong and that their misconceptions must be replaced by the “correct”
conceptions. In order for a new conception to become operative in a student’s thinking, it must be related
to others that are already there. A simple and efficient way to do this is when the new structure is built out
of elements with which the students are familiar. Further, students need to be shown that there are elements
in their experience that can be related differently from the way they habitually relate them. To make such
changes desirable to students, they must be shown that the new way provides advantages and thinking that
reaches beyond passing exams and getting good grades. He finds it counterproductive to dismiss a student
as wrong and then show them the right way to proceed. He believes this disregard to a student’s effort
inevitably kills motivation. He recommends reviewing the method the student uses to give the teacher a
clue to a conceptual connection that is either missing, or, perhaps, has been attained and can be built upon.
On their use of journal writing, Ellerton and Clements (1990) note a common thread in the research that
people of all ages who are studying mathematics have difficulty in reflecting on the deeper meanings of
what they are studying. If teachers of mathematics provide sensitive and constructive advice to students
on how to make journal entries more reflective, if they are able to find the time to read and respond to the
entries, and the learners make regular and thoughtful entries, then they find the journal writing process
to be very valuable.
Southwell (1988) carried out a study to investigate the relationship between experience and reflection
on that experience. In particular, she studied third year teacher education students. Southwell states that
one of the critical issues in learning mathematics which has not been adequately covered is the balance
between theory and practice or the interplay between experience and actual acquisition of concepts.
Techniques devised to enhance the reflective process need to be applied to mathematical problem solving
and to mathematics education research. She states that although there is a general acceptance of the
necessity for reflection in order for learning to be effective, not many have attempted to define or describe
just what reflection is or to develop reflective strategies.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 261

Waywood (1992) reports on the experimental use of writing in secondary mathematics classrooms. The
experiment included about 500 students from diverse socio-economic backgrounds and had been running
for 4 years. There were many students for whom English was a second language. The experiment was a
process of refinement through action and reflection.
While there are many different ways to keep a mathematics journal, the heart of the process used by
Waywood (1988) is the use of prose to review, reflect on, and integrate concepts. His idea has been largely
encouraged through the “Writing across the curriculum movement” and then in the “Writing to learn”
movement. However, Waywood notes, these movements have not had a large impact in the mathematics
classroom. Yet it is clear that mathematics is richer than just a collection of formulas, and writing must
be for us a way of accessing this richness.
The journal writing was analyzed by a rubric with the labels “recount,” “summary,” and “dialogue.”
The point was not to just recognize differences, but to determine how these styles of writing relate to the
learning of mathematics. The teachers’ experience of the students who wrote journals led to the following
hypothesis: “The mode of journal writing reflects a stance towards learning on the part of the student.”
(p. 35). His assertion is that there is a link between different dispositions that students have towards
learning and the different ways that students have of organizing their writing.
In the recount mode, the key feature is a reporting of what happened. It is a passive observation. The
summary mode has the essential quality of summarizing the codifying of content. It may be for preparing
for an exam or more generally to gain a mastery of content. The dialogue mode has the basic feature of
a “to and fro” or interaction between several different ideas. This signals a more creative stance towards
knowledge, knowledge being what is created or recreated.
The student understanding of the journal writing task proved to be essential. A first rationale for the
journal tasks was that they seemed to cover what intelligent learning is about; the tasks included summary,
questioning, collecting examples, and discussing. However, experience with students trying to carry out
these tasks led to the view that these words were only tokens for the changing interpretations in students’
minds. Each of these tasks was interpreted differently, depending on the students’ stance towards learning
and their experience with journal writing.
Students intend to do what is required of them, so what they do reflects their understanding of what
they think they were asked to do. Students did make a shift from happenings to relations.
In this experiment, teachers became more aware of themselves as communicators and paid more
attention to how the learning was organized. They became aware of and began to address elements of
learning that had not traditionally been part of the mathematics instruction.
Waywood offers four conclusions based on this experience. They are:
1. A clear understanding of what is intended by “keeping a journal” must be communicated.
2. Class and homework time has to be given to supporting the use of the journals.
3. The journal work must be seen to be as highly valued as more traditional aspects of mathematics
learning: journals need to be assessed and reported on.
4. Ideally journals need to be introduced at the level of departmental policy.
Richards (1990) reported asking her students to keep a record of what language, reading, writing,
speaking, and listening, they used in 1 week’s time in their mathematics class. She also had a colleague
observe the class, focusing on the modes of communication that were evident. There was a surprisingly
large range of language used. For example, from writing alone, instances of the following types were
observed:
262 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Summaries: of findings, processes used; learning done.


Translations: of definitions, information; concepts and how they are applied.
Definition: of terms used; mathematical areas.
Reports: on an area of mathematics; their work and what they’ve done.
Personal writing: feelings; conversational reports and responses; letters of response.
Labels: for diagrams accompanying explanations; numerical representations.
Instructions: for solving a problem (steps involved).
Notes: from books; from other children’s books; from peers’ tutor sessions; from teacher
tutor sessions; ideas to follow through.
Lists: of findings, knowledge; words; symbols/terminology; ideas; questions; goals;
content.
Descriptions: of procedures; conversations.

This names only some of the types of writing observed. Richards assesses that the extent of this list
certainly draws attention to the significance of language factors in mathematics learning. While the
commonly held view is that mathematics is the least dependent on language factors, it could be argued
that in learning mathematics, students experience the most difficulty.
Richards asks an important question about how students are able to learn to encode their mathematical
experiences. Who tells them how to write mathematics? Examples of student writing that appear to be
literally incorrect, but looked at from an intentional point of view actually have meaning, occur with
considerable frequency in her first year mathematics classes. Unfortunately, Richards notes, it is her
experience that students who are sloppy are most often those who do not have or cannot express, a
conceptual understanding. They are also most often the students who make mistakes; they appear to us
not to be rigorous.
Similarly, Davis and Jones (1990) write that we believe that students’ developing sense of mathematical
rigor goes hand in hand with their ability to use mathematical language precisely, consistently, and
creatively. In other words, we believe that the development of rigor is conditional upon a consistent
language that is capable of expressing it. Conversely, the need for an appropriate mathematical language
seems to grow out of a need to express and to be rigorous.
The experiences of these researchers were taken into consideration in the development of these writing
assignments and how the writing experience would be carried out.

3. Methodology

3.1. The participants

This study involved 25 calculus I students at a public 4-year college located in a large metropolitan
city for whom this was the first calculus class. Eighty percent of these students had a language other than
English as their primary language, including Cantonese, French, Hindi, Mandarin, Russian, Spanish,
and Uzbek. Therefore, the writing component was really two-fold for this particular group. While the
primary goal was to focus on communication between the professor and students that would encourage
a reflective process on the part of the students and provide them more personal attention to their needs,
there was also a goal of promoting the use of English. The college presently has a “Writing Across the
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 263

Curriculum” component and these kinds of activities are in the process of being incorporated throughout
a large number of classes.

3.2. The research question

Can writing exercises in an introductory calculus course promote and/or foster reflective abstraction?

3.3. Development of the writing assignments

In this study, reflective abstraction is a mechanism for the isolation of particular attributes of a math-
ematical structure that allows the subject to construct or reconstruct knowledge that is new; that is,
knowledge not previously known. A feature of reflective abstraction is that it clarifies and organizes
logico-mathematical experiences in such a way as to recognize both nuances and broad generalizations
among them. Any new constructions will be associated with knowledge the subject already has.
With this in mind, the questions were designed so that, in some cases, they were phrased in order to
focus on common properties among different objects in order for students to construct or reconstruct
their knowledge. Based on Mason’s (1989) studies using what he describes as the “shift of attention” and
Thurston’s (1990) article, where he illustrates the dependence between constructive mental activity of the
student and the attention the student has on those structures which are to form part of the abstract concept
while being drawn away from those which are irrelevant. In addition, a common theme throughout the
literature was the connection between the distinguishing the common properties among a group of objects
and abstraction. Thurston (1990) emphasized generalization as a principal incentive for abstraction. With
this in mind, some questions begin with specific cases and then move to more general cases in order
to invite attention to possible generalities. Any indication by the students in their writing of making
generalities, pointing out nuances, discussing a new construction or a reconstruction was considered
evidence of reflective abstraction.
These are certainly not the only possible questions. This process is quite different than any kind of
assessment typically used in a mathematics class. I would argue that it is a more difficult process used
here. In addition to having students answer questions correctly in the form of solving math problems,
they are also expected to explain, discuss and generalize their answers. There is a process of learning and
refinement also for the professor in developing these types of questions.

3.4. The writing assignments

While the students were given seven formal writing assignments that were related to their mathemat-
ics work, they were also given a “math biography” assignment the first day of class in order to set a
tone about the importance and also to familiarize them with writing. The math biography and the first
formal writing assignment which asked them to comment on their computer lab (as well as a few other
points) were written so that there could be, basically, no wrong answers possible. The point of these two
was to introduce the students to writing and allow for the professor to give plenty of positive feedback
to whatever was written. From there, each writing assignment became a bit more conceptual than the
previous one. Also, there was less structure about specifics for each question, requiring the student to
interpret the question, and made it open to that interpretation. This allowed the students to write about
any connections they had and it also made it impossible for a student to copy another student’s response,
264 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

since they were all so unique. The writing assignments as given to the students are listed in the following
sections.

3.4.1. Math biography assignment


Your first writing assignment is your own math biography. Please write several paragraphs about your ex-
periences with mathematics in the past. What was your school system like? Where did you go to school be-
fore here? Do you have a favorite teacher? How about a least favorite teacher? What was your favorite math-
ematics class or subject? Please list all the mathematics classes you had in high school as well as any college
mathematics classes you have previously completed. Why are you taking this class? What is your major?
What are your plans for the future? Please feel free to include anything else that you feel is important.

3.4.2. Formal writing assignments


#1 (a) How is the work we are completing in the classroom related to work completed in the Maple
Lab?
(b) What is the relationship between calculus and the graphing work we have done in the lab?
Instead of using algebraic solutions, we are plotting graphs, why?
(c) Is there anything that we have done so far that is not clear to you?
(d) What was the most interesting thing we have done so far? Have you learned anything?
(e) Generalize the translations we have been doing to any function. Why does translating work
for any function?
#2 (a) Explain what a function is and give examples. Explain the different ways a function can be
represented.
(b) Explain what a piece-wise defined function is and give at least one example.
#3 (a) Given the following two composite functions, determine which part is u(x) and which part is
f(u). √
y = (x 2 + 3)4 y = 1/ x − 2
(b) Write one or two paragraphs explaining exactly how you recognized which part is u(x) and
which part is f(u) for these two composite functions.
(c) Write instructions for a friend how “in general” one determines u(x) and f(u) for any
composite functions.
#4 Write one or more paragraphs on the significance or meaning of the following limits. You may
use examples or definitions in your descriptions if you desire, but you must explain the limits in
your own words. A definition copied from the book will not be accepted unless it is clearly
explained in your own words.
(a) limx→1 2x + 1 = 3
(b) limx→4 f (x) = 7
(c) limx→a f (x) = L
#5 Given the following, answer the questions below:
limn→∞ an = L limx→a f (x) = L
(a) Discuss the similarities and/or differences of these two kinds of limits.
(b) What do these limits signify, in your own words?
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 265

(c) Give examples for when each type of limit does exist and when it does not exist and explain
why.
#6 Which type(s) of function(s) require the following rules in order to determine the derivative
function? Give examples and explain why each rule applies.
(a) Chain Rule (b) Product Rule (c) Power Rule (d) Quotient Rule
#7 (a) Give several examples of functions and graphs of functions which contain a point or points
that have a horizontal tangent. Specify these points and explain why they have horizontal
tangent lines.
(b) Explain why the derivative is equal to 0 at these points listed in 7(a).
(c) What is the connection between the derivative being equal to 0, the extrema points, and the
instantaneous rate of change? Explain in paragraph form, use examples, and write in your own
words.
(d) Give at least three different examples of functions with their graphs that contain at least one
point that is not differentiable. Specify these points and explain why they are not differentiable.
(e) How is continuity of a function related to the derivative? Explain in paragraph form and give
examples.
(f) Does continuity imply differentiability? Does differentiability imply continuity? Explain
your answers.
(g) What else can you say about any of the topics above?

3.5. The procedure

The math biography was completed on the first day of class, in order to set a tone on the significance
of writing. Three of the writing assignments were completed in class and the others were done outside of
class. Each of the math biographies received written feedback and were returned the following class. All
writing assignments were returned within two classes. In addition to the seven formal writing assignments,
there were many informal writing assignments. The formal writing assignments were kept in a journal
which was collected after each assignment and reviewed.
von Glasersfeld (1996) argues that student responses should not be simply dismissed as wrong by the
professor and the student shown the right way to proceed. He suggests reviewing the method used by
the student in order to get a clue of the conceptual connection in order to build on that. Bishop (1985)
also discussed the importance of allowing students to make connection of their own. With these per-
spectives in mind, the written comments and questions to each writing assignment had several goals.
These included showing counter-examples to incomplete reasoning, pointed questions to bring another
perspective, questions to get them to expand further, and encouragement for complete well thought out
answers as well as simple corrections for small errors. They were never told an answer was wrong, but
rather asked to explain further, or directed to consider an example which would allow them to draw their
own conclusions. In some cases, students were asked to rewrite or write further and return the journal.
This would then be commented on a second time and returned. All of the comments and question written
by the professor were with the purpose of keeping the focus on the important properties of the problem,
while allowing flexibility with the descriptions and connections the students were making. Ellerton and
Clements (1990) had noted that if mathematics teachers provided advice on how to make journals more
266 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

reflective, if they read and respond to the entries, and the students make regular and thoughtful entries,
the journal writing process was very valuable. This model was therefore applied here.
As mentioned earlier, the main goal was to promote reflective abstraction. Experimentation with writing
in calculus was piloted by the author earlier, but not so extensively. As Richards (1990) also found in her
studies of writing in mathematics, in these pilot studies, it was encountered that when students were not
able to express themselves clearly in writing on a particular concept, they also were not able to work with
the concept as an object. It was also found in these previous trials that writing was a very personal way to
communicate with the students and that they responded positively. Both in the previous trials and in this
study, students would often approach the instructor to discuss what had been written in response to their
writing, even if it had merely been an underlined sentence or a stray mark. They were very attentive to
anything written directly to them. It was decided this would be a productive way to get their attention and
to document their progress (or lack thereof). Speaking to them, answering questions in class, or even going
around to their desks while they worked in class did not have the same mindful responses from them.

3.6. The assessment

The writing assignments were worth 5% of their grade and it proved to be enough to register considerable
importance with them, since all of the students handed in all of the writing assignments. In addition, there
was a weekly Maple Lab assignment. Students completed interactive lab notebooks which included a
writing component as well. This study, though, focuses solely on the journal writing.
The writing assignments were designed to invite reflective abstraction; to find generalizations and
differences; to organize classifications; and to use all of these to construct or reconstruct knowledge.
However, given that these students in particular come with such a broad range of background experiences,
the comments to their writing were given in such a way as not to dismiss an answer given in any way, but
to broaden it, if possible. The responses were given in an individual way so that students would feel free
to use whatever references they had and to describe their understanding in a way that was natural to them.
That included not marking grammar or misspellings. If a student obviously put very little thought into
what was written or had copied something out of the book, they were asked to rewrite the assignment, but
that was rare. The goal was to get the students to discuss the calculus concepts in question in an abstract
manner and to allow them to relate these concepts to their previous constructs.
For the purposes of this study and based on the previous works, reflective abstraction is a mechanism.
This mechanism can be used to isolate particular characteristics of a mathematical structure. This allows
the student to construct or reconstruct knowledge that is new; that is, knowledge not previously known.
Reflective abstraction may be used to clarify and organizes logico-mathematical experiences, not only
by their nuances, but also by recognizing broad generalizations among them. Therefore, if students
wrote about new constructions, demonstrated by recognizing nuances or generalities, reconstructing with
previous objects or schemas, or organizing the common properties of a group, this new knowledge was
considered a product of reflective abstraction and that reflective abstraction had taken place.

4. Analysis

There were initial responses (the first response given to a question) clearly demonstrating that students
had used reflective abstraction in that they were generalizing and grouping specific attributes of functions
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 267

or focusing on particulars. Some made references and analogies to previous experiences. It is safe to
assume that any knowledge of calculus and function as related to calculus was a new construction since
none of them had studied calculus previously. Other demonstrations of reflective abstraction came when a
student rewrote an answer and articulated a reconstruction of what had previously been written. This may
be due to a comment that caused a reflection on different aspects or it may have been due to something
worked on in class or it could have been from something completely different. The point being that the
writing, and the comments to their writing, initiated some student reflection as demonstrated in their
responses.
Given the individual nature of reflective abstraction and the construction of knowledge, this particular
type of study does not call for quantitative results. Rather, a description of some of the typical types
of writing that students offered follows, as well as some very striking responses. Examples of writing
that demonstrate an occurrence of reflective abstraction are given below. In addition, student work that
establishes difficulties is also included to show their usefulness. The writing served the purpose of a
conduit for information between the students and professor, alerting the professor to misunderstandings
or trouble points for students. This also proved to be very helpful.
Student responses are given verbatim, with any spelling or grammatical errors. Any underlining, paren-
thesis, or quotations in the student responses are taken directly from the students’ own writing. Comments
in [. . . ] are offered by the author if it was deemed that the student writing would not be understood by
the reader.

4.1. First responses demonstrating reflective abstraction

#1 (a) How is the work we are completing in the classroom related to the work completed in the
Maple Lab?
(b) What is the relationship between calculus and the graphing work we have done in the lab?
Instead of using algebraic solutions, we are plotting graphs, why?
(c) Is there anything that we have done so far that is not clear to you?
(d) What was the most interesting thing we have done so far? Have you learned anything?
(e) Generalize the translations we have been doing to any function. Why does translating work
for any function?
Chi: How does translating work? It work the same for all function. For example, within the function
if we subtract, the graph shifts to the right. This is because the x that used to give a certain y now
has moved to the right by whatever subtracted. It has to be bigger to get same effect. The same is
true for adding only x has to be smaller now. If we add outside the parenthesis, this just makes y
bigger, x stays same. So, graph goes up. The same is happening if we subtract outside, the y gets
smaller. When we multiply a number, the graph will become narrower than the original. because
if we multiply, the function will shrink since y gets much bigger for the same x. Therefore, the
graph becomes narrow unless it is multiplied by a fraction then gets wider since y is smaller.
Negative signs change y to negative or positive depending original. So, any function, these are
always true.
Chi has been able to generalize the translation properties across functions as a class. He can reason why
the properties hold and has developed an inter-relation between the algebraic process and the geometric
interpretation.
268 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

#2 (a) Explain what a function is and give examples. Explain the different ways a function can be
represented.
(b) Explain what a piece-wise defined function is and give at least one example.
Mary: Believe it or not, all this material is related. Mostly by the fact they all have to do with function
— even though they be a little different than the next — and by the fact that in each of them
our main goals are to find their domain and range — and to graph them accordingly. Piecewise
functions are just defined different in different parts of the domain. All of them have each value
in domain going to just one value in range. Otherwise, if x goes to more than one range value,
we can’t tell which y we are talking about, so it cannot be a function if this happens.

Mary has been able to connect that all the functions have commonalities and that the domain and range
of these functions, while different, are each related to the graphs of those functions. She also writes very
clearly about how a function is defined.
Maria: Function can be in lots of different forms. A point is a function. Or a graph, more than one
point or an equation. It’s a function if there is only one y for each of the x’s. So a circle is not
a function since there is two y’s for one x sometimes. Piecewise functions are just the same,
except the domain is given and more than one equation is used. But only a function if one y for
each x. Piecewise are sometimes not functions, too if there is x that goes to 2 y’s or more.
Maria is able to articulate clearly the definition of function, how this can manifest itself and when the
function relationship does not hold. She also sees the piecewise defined function as an extension of her
notion of function.
#3 (a) Given the following two composite functions, determine which part is u(x) and which part is
f(u). √
y = (x 2 + 3)4 y = 1/ x − 2
(b) Write one or two paragraphs explaining exactly how you recognized which part is u(x) and
which part is f(u) for these two composite functions.
(c) Write instructions for a friend how “in general” one determines u(x) and f(u) for any
composite functions.
Leon: To determine the u(x) and f(u) for any composite function, you first extract the inside function
(u(x)). To do this, which you would do by taking the part of the function which could stand
on its own as a substitution for x within a function of f(x). After extracting that part of the
function and placing it as u(x) the outside function (f(u)) would be the entire function with the
u(x) part substituted by “u.” That way you can tell which function has been composed in the
other function.
Leon is able to unpack composite functions into separate functions and to see the operator as a function
itself. He speaks of u(x) as a substitution for x in f(x), categorizing this information in his schema for
functions.
Xue: u(x) always inside of a f(u) function, We can see the u value is in the square root, prendency
[parenthesis] and f(u) is to put the u(x) inside of its function, which is function inside of function.
[. . . ] Each one [u(x) or f(u)] could stand by itself and be functions.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 269

Xue is building groups of situations in her schema of function, now adding this description of composite
function.
Gilda: For example, in a fruit, the seed is always in the “inside” of the fruit, so if you’re looking for
a seed you would have to keep biting into the fruit in order to obtain the seed. In this case f(u)
would be the apple and u(x) would be the seed. The only way you would get u(x) is if you got
f(u) = the apple & bit it.
Gilda is taking her experience with composite functions and connecting it with her own experiences
from the past. She is building a schema for composite function which has a representation of the
seed inside the fruit. Gilda’s complete answer is given in Figs. 1–4 along with the comments of the
professor.
#4 Write one or more paragraphs on the significance or meaning of the following limits. You may
use examples or definitions in your descriptions if you desire, but you must explain the limits in
your own words. A definition copied from the book will not be accepted unless it is clearly
explained in your own words.
(a) limx→1 2x + 1 = 3
(b) limx→4 f (x) = 7
(c) limx→a f (x) = L
Sumita: As the difference between x-values and 1 become less then the difference between f(x) and 3
also become less.
Sumita has developed a dynamic interaction between the domain and the limit and uses this in her personal
definition.
Hilda: x is moving towards a number. When this happens, the y values are also moving towards a
number. That number is L. How close do the numbers get? x can get infinedtly [infinitely] close
to the number 1 or 4 or a. Then y values move infinedtly close to L. x or y can sometimes
touch the numbers, but only if the numbers are inside the domain or the range. Sometimes the
numbers are not and sometimes they are, depends on which function. Sometimes there is a hole
or something at the x value or at the y value, but still has limit. Depends.
Hilda writes about her connections between the limit, the domain and the range. She appears to have
some flexibility with the notion on limit, depending on which functions she is working with.
#5 Given the following, answer the questions below:
limn→∞ an = L limx→a f (x) = L
(a) Discuss the similarities and/or differences of these two kinds of limits.
(b) What do these limits signify, in your own words?
(c) Give examples for each limit of when it does exist and when it does not exist and explain
why.
Lin: If ε is big the M will happen fast. If ε is small then the M will happen slow.
Lin also has a dynamic interaction here, between ε and M, and has incorporated her idea of fast and slow
into her schema of the sequence limit.
270 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Fig. 1. Gilda: Question 3.

Babul: In both the domains are approaching values. As x or n approaches a or infinity, the distance
between y values and L gets smaller if the limit is true. If the distance between y and L gets
bigger, or jumps around, the limit is not true. This is true for both limits. One is sequence, n
just gets bigger. It’s like the horizontal tangent.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 271

Fig. 2. Gilda: Question 3.

Babul has been able to make some important generalizations and connections between the two limits. He
is constructing his schema for limits and including the horizontal tangent.
#6 Which type(s) of function(s) requires the following rules in order to determine the derivative
function? Give examples and explain why each rule applies.
(a) Chain Rule (b) Product Rules (c) Power Rule (d) Quotient Rule
Maya used y = (3x − 2x 2 )(5 + 4x) as an example for the product rule. She finds the derivative using
the product rule and then writes:
272 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Fig. 3. Gilda: Question 3.

Maya: It is possible to simplify the equation then take the derivative by using the power rule. However,
it makes the problem more complicated, and therefore, we use the product rule.
She then gives the example of (3x − 2)/(2x − 3) and finds the derivative using the quotient rule. She
then writes:
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 273

Fig. 4. Gilda: Question 3.

Maya: It is possible to write the function in the form of uv, such as (3x − 2)(2x − 3)−1 then use the
product rule. We can use both rules. [. . . ] But it is necessary by looking at the functions and
determine which rule will solve the problem easily and in short time.

Maya is able to determine the commonalities among functions and writes them in their various forms of
products or quotients. She can then use this to determine which is most efficient. She is able to make the
shift of attention and use whichever classification is suitable.
274 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Shiela: We can say that the power rule is the simple case of the chain rule. [. . . ] There is no basic
difference between chain rule and power rule. Both methods are used to take the derivatives
and results are the same. However, for some functions by using power rule we can save time.
In the same way, for some functions by using chain rule we can make the problem simple and
easy.
Shiela has constructed her schema for the derivation rules and for functions and is able to draw on these
as is most efficient for the situation. She has made connections between the chain rule and power rule
and can group them in the same category or separate them, as needed.
Carol: All these rules can be used for the same function sometimes. For example, I can use the chain
rule or the power rule for x 3 + 6x 2 since I can just let x = the composed function, the one on
the inside that’s y = x. But that’s stupid since it takes more work. But you could if you wanted
to any time. But if I have y = (3x + 4)5 then only the chain rule works unless I multiply it all
out and that would be stupid this time. When we have a function dividing a function, I use the
quotient rule but sometimes I use the other ones too. So if I have sin(x)/3x2 , I could use quotient
or product rule whichever I want. They both work. Sometimes I use one rule “inside” another
rule. Like if I have x sin(x4 ) I use the chain rule inside the product rule. At first, I was confused
about what rule to use then I see it not that hard.
Carol is able to distinguish between the different functions to see that the rules can be used interchangeably
in certain situations. She also has realized that some rules are more efficient in certain circumstances.
She is able to coordinate the different functions with the different rules and make a decision based on her
understanding of the different rules.
Maribel: The chain rule for example uses the equation of d/dx[f (u)] = f  (u) u and like all the
other above rules can be used with composite functions. [. . . ] Each rule can be used with
each of the two types of functions I have mentioned [composite functions and trigonometric
functions] because they can always be combined and use more than one rule to solve an
equation. However if the function, for example, is just a function with an exponent you know
to use the power rule. [. . . ] The relationship between the two rules (power and chain) is that
the power rule is a special case of the chain rule.
Maribel is repeatedly able to generalize all the rules across all functions. She is able to group them together
if needed, or use them individually, using the characteristics of the particular function to decide.
Mia: As you see in above [Her example uses both the chain rule and the product rule to find the
derivative of the same function], you get same result. But not every equations we can use both
chain rule and power rule to get same result. Some equations have to use only power rule, product
rule, or quotient rule. Chain Rule and Power Rule are similar and differ in some ways. If you use
chain rule in some equations, the algebraical process might be longer or complicated but if you
use power rule, you can get the answers easily. But no matter which rule you use, you end up
with same result.
Mia has generalized the rules and found that they can be interchanged, but it may not always be useful to
do so, depending on the function in question. While she is building her schema for derivative, she does
not express that the power rule is necessarily a special case of the chain rule.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 275

#7 (a) Give several examples of functions and graphs of functions which contain a point or points
that have a horizontal tangent. Specify these points and explain why they have horizontal
tangent lines.
(b) Explain why the derivative is equal to 0 at these points listed in 7(a).
(c) What is the connection between the derivative being equal to 0, the extrema points, and the
instantaneous rate of change? Explain in a paragraph form, use examples, and write in your
own words.
(d) Give at least three different examples of functions with their graphs that contain at least one
point that is not differentiable. Specify these points and explain why they are not differentiable.
(e) How is continuity of a function related to the derivative? Explain in paragraph form and give
examples.
(f) Does continuity imply differentiability? Does differentiability imply continuity? Explain
your answers.
(g) What else can you say about any of the topics above?
Bella: The connection between the derivative being zero, even though it doesn’t imply that there is a
min or a max; however, when there is a min or a max it does imply that the derivative could be
zero (or undefined). And therefore the instantaneous rate of change (velocity) is also found by
what the tangent line is. Therefore if the tangent line is zero, the instantaneous rate of change is
also zero, since the derivative equals the instantaneous rate of change.
Bella broadly discusses the relationships between these concepts, making their connections, and reasons
with the conditions of their existence.
Maria: So, continuity doesn’t imply differentiability, however differentiability does imply continuity
since differentiability means that you can have a tangent at that point and therefore has to be
continuous.
Maria discusses the relations between continuity, differentiability, and the tangent line and points to their
geometric interpretation.
Henry: The derivative is equal to zero at these points, because the derivative is defined as the change in
(y) with regards to (x). Since the line is horizontal at these points, (tangent line) then there is no
change in the (y) value. There is zero change in the (y) so the derivative is zero. The derivative
at the maximum and minimum points is equal to zero as the tangent lines at these points are
horizontal, therefore no instantaneous rate of change in (y).
Henry has constructed his understanding of derivative based on the relation between slope, change in y
to x, with horizontal lines and instantaneous rate of change. He is demonstrating a cooperation of these
ideas to form his schema of derivative.
Gilda: Thus, you see that continuity does affect differentiability b/c if a function is discontinuous
at a certain pt. How can we find the derivative? Then I’d say that continuity doesn’t imply
differentiability but differentiability does imply continuity.
Gilda is discussing the relation between continuity and her schema for derivative to reason that her
understanding of the derivative necessitates continuity.
276 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Sarah: Any thing that has a motion, in order to achieve the highest or lowest speed it has to be stop at
certain point. In the same way, when a function is increasing then decreasing with in an interval
it has to be zero at certain point.
Sarah is using her previous knowledge of motion to construct her understanding of extrema values.

4.2. Reflective abstraction demonstrated in a rewritten response

Some of the best documentation of promoting the reflective abstraction was in asking students to
rewrite a response. Their original responses would be commented on with questions, examples, and
counter-examples. They would then be asked to read what was written, think it over, and rewrite their
response. Therefore, the comparison of the original writing and the rewritten response would allow for a
demonstration of a reconstruction by the student.
#3 (a) Given the following two composite functions, determine which part is u(x) and which part is
f(u). √
y = (x 2 + 3)4 y = 1/ x − 2
(b) Write one or two paragraphs explaining exactly how you recognized which part is u(x) and
which part is f(u) for these two composite functions.
(c) Write instructions for a friend how “in general” one determines u(x) and f(u) for any
composite functions.
Hans: (original writing) In the above equations u(x) is the inside function and is the direct value of x.
In the first equation (x 2 + 3)4 , the direct value of x is found from (x 2 + 3). f(u) is the outside
function and dictates what is done to the inside value.
Hans: (rewritten version) In the equation y = (x 2 + 3)4 , the u(x) value is x 2 + 3 which is the inside
function. The u(x) determines what is first done to the equation. The f(u) is u4 and is the outside
function and directs what is done to the inside function.
[. . . ] To do these functions first determine the inside function which will direct the first
operation to be carried out on the equation. The outside function is next determined. This is the
operation that is done to the inside function which condenses the overall equation.
Hans is able to discuss the composite function as two separate functions. In the second writing, he is
able to describe them as “condenses the overall equation” and describe the interactions between the two
functions.
Susan: (original writing) This means we can plug in any positive or negative numbers or 0 in the
equation, consequently we get the value of y which is the range.
Susan: (rewritten) However, the range is [−4, ∞). That means we can get any values that are between
[−4, ∞). But we cannot get any y-value that is less than −4. The range will be whatever comes
out of the function, depending on which function it is.
In Susan’s original response, there was a vagueness to the idea of range, stating only that it is the value of
y. In her second response, she is able to make a stronger statement about the range and to actually find it.
Carol: (original writing) The u value is contained and the other is not. You need to look for each
one.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 277

Carol: (rewritten) One function has been input to other function. So what used to come out of the u
function now goes into the other function. It’s like machine with another little machine in it.
Now the little one makes the bigger one run. One function is folded into the other one and even
though they were two now that they are one.
In Carol’s rewritten work, she is showing that she is developing the idea of composite function from
something very limited to discussing the domain and range (using the term input). She is also making
connections with what is familiar to her with comparing a function with a machine.
#4 Write one or more paragraphs on the significance or meaning of the following limits. You may
use examples or definitions in your descriptions if you desire, but you must explain the limits in
your own words. A definition copied from the book will not be accepted unless it is clearly
explained in your own words.
(a) limx→1 2x + 1 = 3
(b) limx→4 f (x) = 7
(c) limx→a f (x) = L
Linc: (original writing) By limit I mean, the value that the y values will get closer and closer to, but
never actually equal that value. [. . . ] This means that while the x value of f(x) get closer to the
value 4 from the right and the left the y values (f(x)) get closer to the value 7 but will never
actually be equal to it.

After making several pointed questions and giving him examples, in the next writing assignment, Linc
tries to compensate and points out repeatedly, four separate times, that y may actually be equal to L or
may not.
Linc: (rewritten answer) However in “b” when the y values get closer to the limit or even equal the
limit the x values, get closer or may equal a specific value depending on the function. (He repeats
this point several times in his answer.)

4.3. Examples of misunderstandings as written by students

#5 Given the following, answer the questions below:


limn→∞ an = L limx→a f (x) = L
(a) Discuss the similarities and/or differences of these two kinds of limits.
(b) What do these limits signify, in your own words?
(c) Give examples for each limit of when it does exist and when it does not exist and explain why.
Rajendra: They are similar because both functions have the same limit.

This comment was interesting since the student was interpreting that the limit of the sequence and the
limit of the function were both approaching the same value, since they were both approaching L.
#6 Which type(s) of function(s) require the following rules in order to determine the derivative
function? Give examples and explain why each rule applies.
(a) Chain Rule (b) Product Rule (c) Power Rule (d) Quotient Rule
Racquel: The chain rule is usually used in composite functions.
278 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

The key word here is “usually.” Some students would start off by making sweeping generalizations in
their writings and would find that this often would not hold. They had a tendency to want to generalize
broadly as much as possible. In response, they would be given counter-examples to show that their rule
was not universal. Some students would then go in the other direction and would write words such
as “usually” or “most of the time” or “sometimes” in cases where the generalizations did hold more
broadly.
Hector: For the function y = x 2 + 2x, the chain rule could not be used as there was no part of the
function that could be defined as the inside or outside function.
There were a number of similar comments from students who did not make the connection between the
chain rule and the power rule.
Joan: The different between the chain rule and the power rule is the chain rule has to find the derivative
of the inside and outside function and the power rule don’t.
Joan’s response was after correctly finding the derivative using both methods for the same function. This
could be an example of someone who is not able to clearly express themselves.
#7 (a) Give several examples of functions and graphs of functions which contains a point or points
that have a horizontal tangent. Specify these points and explain why they have horizontal
tangent lines.
(b) Explain why the derivative is equal to 0 at these points listed in 7(a).
(c) What is the connection between the derivative being equal to 0, the extrema points, and the
instantaneous rate of change? Explain in a paragraph form, use examples, and write in your
own words.
(d) Give at least three different examples of functions with their graphs that contain at least one
point that is not differentiable. Specify these points and explain why they are not differentiable.
(e) How is continuity of a function related to the derivative? Explain in paragraph form and give
examples.
(f) Does continuity imply differentiability? Does differentiability imply continuity? Explain
your answers.
(g) What else can you say about any of the topics above?
Dani: When a function is continuous, it definitely has a derivative. For instance with polynomials:
y = x 3 − 2x + 15, the function is continuous everywhere, therefore it is differentiable because
there’s no jumps, holes or cusps. If the function is continuous, it must be differentiable because
it is continuous for all the points, and, on the other hand, if a function is differentiable, it has to
be continuous, too.

Dani is stating that she does not have a clear concept of differentiability or continuity or their relationship.
Harold: To be continuous, a function has to have an (x) value that defines a (y) value on the graph
of the function at every point. If the function is discontinuous then the (y) value at that point
does not correspond on the graph to an (x) at this point. There therefore can be no change in
the (y) value if there is no (y) value to change, at these points. The derivative therefore cannot
exist.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 279

Harold has constructed a geometric interpretation of continuity and his understanding of the points that
make a graph with his schema on derivative to bring together their relationship. However, he misses the
function that can have both x and y in the domain and range and still be discontinuous.

5. Conclusions

There was an interesting trend in reviewing the journals: there were more indications of reflective
abstraction with each writing assignment. In other words, the journals as a whole contained more examples
of students classifying and organizing information, making connections and discussing relationships
between concepts, generalizing across categories, noticing particulars in certain circumstances while
drawing away in others. It appears that they became better at the process as the semester progressed. I
believe they did and the reasons why are varied.
One reason students demonstrated more reflective abstraction may be that they were better at writing.
Many of them seemed anxious in the beginning of the semester about having to write because English
was not their first language and they thought that they would be graded on grammar and/or spelling.
After several assignments, they realized this was not the case and that what they wrote was the important
issue. They became freer about sharing what they thought and knew better how to express themselves in
the mathematical sense. In addition to developing their talents for writing mathematics, they were also
learning more mathematics in the process. Therefore, they had more structures available to draw on in
their process of constructing knowledge as the semester moved forward.
The writing assignments themselves became more demanding, and required more responsibility from
the students. They were designed to start simply, with the math biography and description of their Maple
Labs. The comments from the professor would ask for more information or to clarify or simply to respond.
This laid the groundwork for explanations and greater thoughtfulness. The goal of each writing assignment
was to elicit a conceptual response from the student. With each writing assignment there was more to
reflect on. The goal was to promote the use of reflective abstraction as a means of promoting greater
understanding and knowledge construction.
The communication that the journal writing provided between professor and student was a very strong
tool for determining student understanding or lack of understanding of particular topics. Each writing
assignment is a diagnostic assessment, in which the professor can focus on the particular results of
each student. This drew the students’ attention. No other activity in class produced such an attentive
reaction. As Bishop mentioned, teaching activities are important in the mathematics class, as opposed to
the teaching methods. While he focuses on spatial activities, writing stimulates the students to use their
personal experience, allows for a steady flow of communication between student and teacher, and gives
a personal aspect to the education of each student.
As Waywood found, it was extremely important for the writing assignments to be commented on and
returned in a prompt manner. First, if students needed to rewrite a section, it was necessary to allow them
time to do so before the next writing assignment. Also, students wanted the assignments returned in a
timely fashion to see how they had done. It would not have made sense to comment if they were having
trouble on a topic and not return it until weeks after that topic was finished in class. Therefore, to keep
it all as part of the learning process, it had to stay within a timely exchange. At points, this was arduous
and time consuming. However, it was not as much work or as time consuming as I thought it would be
and the results made it more than worthwhile.
280 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

There were several strong outcomes from this experiment with writing in calculus.

• The students made it evident that they wanted to answer well and that the writing assignments were
important to them. They responded to the professor’s comments and demonstrated a seriousness of
purpose with their written responses. They were able to give a kind of self-expression that is virtually
impossible in a typical homework assignment or in the classroom setting. This attention to detail
can promote the reflective abstraction process that they need in order to construct or reconstruct the
knowledge for calculus.
• The communication that is initiated through this process of writing was beneficial not only to the
students, but also to the professor. Students were able to have a type of diagnostic analysis of their
own understanding via the comments given by the professor. The professor in turn received a strong
basis for understanding what kinds of knowledge the students have constructed, and which particular
areas needed attention.
• The process of writing, comments, and further writing produced evidence of the process of reflective
abstraction for many of the students. It is not clear, nor could it ever be absolutely clear, if the
questions asked of the students or the thinking they do in order to write a response are the catalysts or
if the writing assignment is a conduit by which they can demonstrate that the process has occurred.
However, in either case, the writing assignment gave them opportunities to articulate their new
constructions.

6. Discussion and recommendations

In contrast to humanities classes, for example, in a typical college mathematics course there is very
little room for discussion. Granted, it can be very difficult to have a genuine exchange about very new
and rigorous concepts. However, students are still expected to be able to discuss such ideas. If there is not
discussion in class, it is not clear exactly how students can learn to reflect, ask questions, and discuss the
mathematics they are learning. In other words, upper level mathematics, in particular, is taught more as
if the professor has the information, and that information is dispensed in a manner that, ultimately, will
make it understandable to the students. However, in a philosophy class, for example, there is a constant
exchange of ideas and discussion of the meaning of what they are learning. Thus, philosophy majors are
known for being able to discourse about what they are learning, whereas mathematics students often have
difficulties articulating what they are learning. I would argue that this is because they never learn how to
articulate these ideas, nor how to discuss or debate these ideas. They may not be able to ask meaningful
questions since they do not know how to put their questions into words. Journal writing offers a conduit
for discussion. It helps students to articulate their thinking and to reflect on what it is they are learning.
Reading through the journals, I noticed that the writing at the end of the semester was more expository,
had more examples, and was better organized than the writing at the beginning of the semester. If such
progress in the way that students can express their mathematical ideas can be noticed in just one semester’s
time, it follows that an even markedly higher improvement might be noticed over one or more years. A
writing component of some sort throughout a mathematics department could produce students that not
only reflect on what they are writing, but who are also better able to use this writing in such important
matters as proof. This constant reinforcement of explaining, justifying, and writing is a very valuable tool
in higher mathematics, not to mention in critical thinking in general.
L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282 281

Bearing all of this in mind, the philosophy of this writing in mathematics study was to promote reflective
abstraction. This would entail, as indicated by Dubinsky and Lewin, a process of cognitive development.
In order to abstract, the students would need to construct or reconstruct their knowledge on a new plane
and become aware of constructs that they were not previously conscious of.
Given the broad agreement demonstrated by the authors listed previously who had written on reflective
abstraction, such cognitive development is not easy to promote. The writing assignments and comments
written in response to the students were developed to promote what Mason refers to as “the delicate shift
of attention” that comes so easily to people who have a strong proclivity for mathematics. Questions were
meant to invite students to generalize and synthesize where needed and to focus on key particulars in
other places. As Tall described abstraction as an isolation of specific attributes and as Borasi refers to it
as a class of situations, all promote a way of clarifying and organizing logico-mathematical experiences
in such a way as to recognize both nuances and generalizations. These authors also agree that any new
constructions must be associated with knowledge the student already has.
In the future, the writing-rewriting model should be further explored. In the end, I found that, while this
model required more from the professor, it produced more satisfying results. It would be useful to have
students respond to the previous comments each time they hand in a new assignment, thus expanding the
communication. One very pleasant discovery in this study was the willingness of the students to write
and participate in the writing activity. They were in general very cooperative and open in their writing.
They gave a very honest effort in carrying out the writing assignments, in that they tried to articulate what
they were learning in a meaningful way. We were able to open the door to communicating mathematics
in a more interactive way.

References

Beth, E., & Piaget, J. (1966). Mathematical epistemology and psychology (W. Mays, Trans.). Dordrecht: D. Reidel Publishing
Company.
Bishop, A. (1985). The social construction of meaning — a significant development for mathematics education? For the Learning
of Mathematics, 5(1), 24–28.
Borasi, R. (1984). Some reflections on and criticisms of the principle of learning concepts by abstraction. For the Learning of
Mathematics, 4(3), 14–18.
Davis, G., & Jones, A. (1990). Reading for comprehension and writing for communication in university mathematics. In: G.
Davis, & R. P. Hunting (Eds.), Language issues in learning and teaching mathematics (pp. 112–119). Victoria: Institute of
Mathematics Education, La Trobe University.
Dienes, Z. P. (1971). Six stages of learning. London: National Foundation for Educational Research.
Dienes, Z. P. (1978). Learning mathematics. London: Mathematical Education.
Dubinsky, E. (1991a). Constructive aspects of reflective abstraction in advanced mathematics. In: L. P. Steffe (Ed.), Epistemo-
logical foundations of mathematical experience (pp. 160–220). New York: Springer.
Dubinsky, E. (1991b). Reflective abstraction in advanced mathematical thinking. In: D. Tall (Ed.), Advanced mathematical
thinking (pp. 95–123). Dordrecht: Kluwer Academic Publishers.
Dubinsky, E., & Lewin, P. (1986). Reflective abstraction and mathematics education: the genetic decomposition of induction
and compactness. The Journal of Mathematical Behavior, 5, 55–92.
Eco, U. (1988) The aesthetics of St. Thomas Aquinas (H. Bredin, Trans.). London: Radius.
Ellerton, N., & Clements, K. (1990). Language factors in mathematics learning. In: Proceedings of the 13th Biennial Conference
of the Australian Association of Mathematics Teachers (pp. 230–258).
Mason, J. (1989). Mathematical abstraction as the result of a delicate shift of attention. For the Learning of Mathematics, 9(2),
2–8.
Piaget, J. (1975). To understand is to invent: the future of education. New York: Viking.
282 L. Cooley / Journal of Mathematical Behavior 21 (2002) 255–282

Piaget, J. (1976). The grasp of consciousness (S. Wedgwood, Trans.). Cambridge, MA: Harvard University Press. (Original work
published 1974)
Piaget, J. (1985). The equilibration of cognitive structures (T. Brown & K. J. Thampy, Trans.). Cambridge, MA: Harvard
University Press. (Original work published 1975)
Richards, L. (1990). Measuring things in words: language for learning mathematics. Language Arts, 67(1), 14–25.
Southwell, B. (1988). Construction and reconstruction: the reflective practice in mathematics education. In: A. Borbas (Ed.),
Proceedings of the Annual Conference of the International Group for the Psychology of Mathematics Education (Vol. 12,
pp. 584–592).
Tall, D. (1988, July). The nature of advanced mathematical thinking. Discussion paper for the Working Group on Advanced
Mathematical Thinking at the Annual Conference of the International Group for the Psychology of Mathematics Education,
Vol. 12. Vezprém, Hungary.
Thurston, W. P. (1990). Mathematical education. Notices of the American Mathematical Society, 37, 844–850.
von Glasersfeld, E. (1996). Aspects of radical constructivism and its educational recommendations. In: P. Cobb, G. Goldin, B.
Greer, P. Nesher, & L. Steffe (Eds.), Theories of mathematical learning (pp. 307–313). Mahwah, NJ: Lawrence Erlbaum.
Waywood, A. (1992). Journal writing and learning mathematics. For the Learning of Mathematics, 12(2), 34–43.
Waywood, A. (1988). Mathematics and language: reflections on students using mathematics journals. In: R. P. Hunting (Ed.),
Language issues in learning and teaching mathematics (pp. 37–38). Melbourne: La Trobe University.

View publication stats

You might also like