Journal of Mathematical Behavior 25 (2006) 334–344

Does mathematical learning occur in going from concrete

to abstract or in going from abstract to concrete?
Wolff-Michael Roth a,∗ , SungWon Hwang b
aUniversity of Victoria, Canada
b Hanyang University, South Korea

Abstract
The notions of abstract and concrete are central to the conceptualization of mathematical knowing and learning. It is generally
accepted that development goes from concrete toward the abstract; but dialectical theorists maintain just the opposite: development
consists of an ascension from the abstract to the concrete. In this article, we reformulate the relationship of abstract and concrete
consistent with a dialectical materialist approach to conscious human activity, as it was developed in the line of cultural-historical
psychology. Our reformulation of development in and through interpretation shows that rather than being a movement from concrete
to abstract or from abstract to concrete, development occurs in a double ascension that simultaneously moves in both direction: it is
a passage of one in the other. In the proposed approach, the theoretical contradictions of earlier approaches to the issue of abstract
have been eliminated.
© 2006 Elsevier Inc. All rights reserved.

Keywords: Abstract; Concrete; Generalization; Dialectical logic; Double ascension

Who thinks abstractly? The uneducated person, not the educated. (Hegel, 1988, p. 575)1
It is necessary to challenge the ideas of proof, abstraction, calculation, low level, and so forth. (Walkerdine,
1997, p. 58, our emphasis)
Who thinks abstractly? Without doubt, most mathematics educators would answer something like “the educated”
or “those who have developed formal operations (reasoning).” It is evident that their responses are precisely the
opposite of what Georg W.F. Hegel articulated in our introductory quote. This difference became salient to us during
an expert/expert study where experienced and highly successful scientists had been asked to interpret graphs from
undergraduate courses in their own discipline (Roth & Bowen, 2003). It turned out that in some instances, half of
them did not arrive at the answers that the professor teaching the course from which the graphs had been culled
would have accepted as correct. Our scientist participants struggled making sense of the graphs, taken to be abstract
representations of natural phenomena (Latour, 1993). However, those scientists who succeeded articulated a lot of
concrete detail and used their lived experiences during their interpretations. To better understand the relation between
the abstract (abstraction) and concrete (concretion), we constructed a detailed case study of one scientist (Eddie)

∗ Corresponding author at: Applied Cognitive Science, MacLaurin Building A548, University of Victoria, Victoria, BC, Canada V8W 3N4.
Tel.: +1 250 721 7885; fax: +1 250 721 7767.
E-mail address: mroth@uvic.ca (W.-M. Roth).
1 We used several non-English works. All translations are ours.

0732-3123/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jmathb.2006.11.006
 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344 335

interpreting a graph featuring the distributions of three types of plants; the graph is used in ecology courses in support
of a theory of adaptation (see Roth & Hwang, 2006).
Our case study articulates a trajectory in the course of which Eddie moved from not knowing the sense of the graph
to the production of a statement that “correctly” described the graph in terms of an abstraction: Differential adaptation
to climate leads to differences in the distribution of plants with different photosynthetic mechanisms. The trajectory
constitutes an episode of learning, where Eddie articulated and appropriated the sense of this graph. Interestingly, our
case study showed how the movement along this trajectory simultaneously is from concrete to abstract and from abstract
to concrete. In our paper we point out that this is not a case of two compatible movements, the first from concrete to
abstract and the second in the opposite direction. Rather, the learning trajectory is equivalent to a double ascension
(some authors use the term ascent instead)—the simultaneous movement from abstract to concrete and concrete to
abstract. This double ascension is precisely the movement along which sense comes to be articulated; and “sense is
the ideality of the sensible and the sensibility of the idea: it is the passage of the one in the other” (Nancy, 2002, p. 49).
The idea (abstract, general) and the sensible world (particular, concrete) no longer are separate—one is the passage in
(not into!) the other. “Sense passes between the two, from the one to the other absence of sense, from the one to the
other truth” (p. 50).
The purpose of this article is to develop and ground a dialectical theory in which mathematical learning is described
as a double ascension the simultaneously and contradictorily moves from abstract to concrete and from concrete to
abstract. In the process, the dichotomy apparent in our title comes to be negated and sublated (a term that has the sense
of making cease, integrate, and being overcome).

1. The problematic

The concept of abstraction, which has recently received a lot of attention from mathematics educators (Dreyfus &
Gray, 2002), really has been an important topic of philosophical and mathematical thinking since the times of the ancient
Greek. Abstraction literally means drawing (Lat. trahere) away (Lat. ab[s]) and frequently is associated with the idea
that it is characteristic of intelligence and higher-order cognition (Ohlsson & Lehtinen, 1997). Naturalist philosophers
define abstraction (or generalization) as a process, where human beings move(d) from simple observation sentences
or terms (e.g., ∃ x(x is raven)) to focal observation categoricals (e.g., ∃ x(x is raven. ⊃ x is black)) to observation
categoricals that describe a number of observations or classes of objects (e.g., ∀ x(x is raven ⊃ x is black) (Lakoff,
1987; Quine, 1995; Russell, 1938; Wittgenstein, 1978). Not everyone agrees, though, with this characterization. There
are suggestions that already in the observation of a new object, the object involves a unity of appearances such that
the apperception of the object as object is “in general the first universal typification—precisely the [typification of
the object] as an object of experience, an object of perception, and the [typification of unities] as a configuration
of objects” (Husserl, 1945, p. 335). In any event, abstraction comes off as one of the highest human achievements,
whereas concreteness is, as in Piaget’s stage theory, characteristic of lower levels of thinking. However, the concept
of abstraction is not without ambiguity and, as our first introductory quote shows, there are contrary formulations that
attribute abstract thinking to the undeveloped person (mind).
A different way of thinking about abstract and concrete suggests that concreteness “is that property which measures
the degree of our relatedness to the object (the richness of our representations, interactions, connections with the object),
how close we are to it, or, if you will, the quality of our relationship with the object” (Wilenski, 1991, p. 198), where,
in our interpretation, the relationship may be to a material object or to an idea. This statement therefore also allows us
to understand the opening Hegel quote in a new way, as “uneducated” may be interpreted to mean that a person has
less rich representations, interactions, and connections with the (material, ideal) object of activity. In fact, dialectical
philosophers generally describe development of scientific cognition to occur in a process of ascending from abstract
to concrete (e.g., Il’enkov, 1982).
The two ways of thinking about the terms abstract and concrete lead us to a contradiction: learning and development
appear to simultaneously constitute a movement from concrete to abstract and a movement from abstract to concrete.
In other words, it is not two concurrent movements one going from abstract to concrete and the other from concrete to
abstract. Rather, it is the same movement (development, learning) that simultaneously goes from abstract to concrete
and from concrete to abstract. To put it in logical terms: if p is the statement “movement is from concrete to abstract,”
then ¬p stands for its negation, “movement is from abstract to concrete.” What our statement therefore suggests is that
p = ¬p. This clearly constitutes a contradiction and is not permissible in classical logic—a statement p and its negation
336 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344

¬p cannot be true (or false) simultaneously; and the law of excluded middle states that there is no third alternative
(“tertium non datur”).
At least one group of researchers articulated a proposal to get us out of this dilemma. Grounding themselves in
the work of the Russian educational psychologist Vasili V. Davydov (e.g., 1990), they suggest that the process of
mathematical abstraction leads from “an undeveloped to a developed form of the abstract in which new features of
the concrete are emphasized” (Hershkowitz, Schwarz, & Dreyfus, 2001, p. 200; see also Tsamir & Dreyfus, 2002).
Whereas we consider this a (theoretical, methodical) move in the right direction, these authors replaced the classical
dichotomous concrete–abstract distinction with a continuum from undeveloped to developed. In the very dialectical
method to which the authors pledge allegiance, this constitutes a move to the worse: “difference as magnitude” “is an
indifferent difference” and therefore cannot represent conceptual understanding and its development (Hegel, 1977, p.
168). The non-identity between abstract and concrete, which affords dynamic transformation of one to the other and
thereby manifests its mutually constitutive nature within the whole, cannot be reduced to a simplified or one-sided
form of identity between the two.
The purpose of this article then is to challenge (as Walkerdine asks us to do in the second opening quote) the
concept of abstraction and thereby to contribute to the discussion of the relationship between the concrete and the
abstract in the context of mathematical knowing and learning. We do so by articulating a dialectical approach of
understanding the relationship between a knowing consciousness (mind) and the object of its activity. We propose the
dialectic of abstract|concrete, which characterizes the unit of practical activity. This unit “sublates” (Hegel’s German
equivalent Aufhebung means suppressing, making cease, and gathering, retaining) and therefore negates both abstract
and concrete. The unit abstract|concrete is not only at the heart of the development of mathematical knowing but also
one of its driving forces. We articulate a dialectical answer to the theoretical problem posed by the idea of a double
ascension from abstract to concrete and concrete to abstract that mathematics educators and educational psychologists
recently articulated. Consistent with the cultural-historical underpinnings of our approach, we trace the history of some
of the fundamental ideas underlying the concept of abstraction in mathematics and mathematics education.

2. Historical perspectives of abstract and concrete

Immanuel Kant points out that we do not abstract something, that is, take something away, but we abstract from
something—the determination of an object of thought—whereby this something obtains the generality of a concept and
thereby is taken up by consciousness (Kant, 1968e).2 The philosopher uses the example of scarlet cloth to concretize
his understanding of the process of abstraction (Kant, 1968d, p. 525). If we think about the red color of the cloth only,
then we abstract from the cloth. If we also abstract from the red color to think about materiality in general, then we
abstract from several determinations, and our resulting concept is even more abstract. Kant points out that abstract
concepts should be called “abstracting concepts,” that is, concepts that contain several abstractions simultaneously
(Kant, 1968c). To put this example into a modern educational context, consider the following example similar to the
ones we recently recorded in a second-grade classroom learning about two- and three-dimensional geometrical objects.
The children are presented with sets of objects, such as

(1)
which they are asked to put into groups. (For simplicity we only include four objects, although the children had many
more; theirs were three-dimensional cones, cubes, pyramids.) At one table, the two children decide on the following
grouping of the objects:

(2)
To achieve the groupings, the children had to have suppressed everything but color, that is, to use Kant’s language,
they have abstracted color from each object and then put those of like color together. In other words, the children have
suppressed other aspects such as shape and size and have retained nothing but the color.

2 German philosophers such as I. Kant and G.W.F. Hegel use the term “Begriff,” which comes from Ger. greifen, seize, grab, hold; the official
translations render the term as “Notion” or “concept.”
 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344 337

For Kant (1968d), “to turn appearances into concept, one has to be able to compare, reflect, and abstract” (p. 525).
The term reflective abstraction therefore is a viable alternative for “abstracting concepts,” because we need to make the
first-level abstractions present again in some form—psychologists frequently use the term representation—and reflect
upon one or more commonalities upon which the new, more abstract concept is based. To arrive at new concepts, that
is, to generalize, one needs to compare, reflect, and abstract, the three logical operations that are fundamental to the
ontological and general determination of any concept.
Taking these ideas as his foundation, Jean Piaget (e.g., 1970) developed his ideas of simple abstraction (from physical
objects of experience) and reflective abstraction (from actions on physical objects), the notions that have been central
to mathematics education research. Thus, one study exemplifies different stages of abstraction as a student goes from
a story problem containing 5 and 12 trout to the question 12–5 = ? and to a formula such as a + b = c, where a and b are
part sets and c is a whole set (Staub & Stern, 1997). The second instance is more abstract (general) than the first, as
the equation can also be used to model story problems about pebbles. Students moving from the story problem to the
mathematical equation engage in a simple abstraction. The third instance is even more abstract (general), as it can be
applied to any problem about addition and subtraction. Moving from doing many addition and subtraction problems to
an understanding of how part sets (a, b) and whole sets (c) are related constitutes a reflective abstraction in a Piagetian
sense. These examples show how much of the work concerning the relationship of concrete and abstract, and on the
process of abstraction (generalization) has been an extension of Kant’s work into the psychology of mathematics
education.
Kant went further, though, laying a foundation for subsequent dialectical formulations of conceptual development.
Thus, “abstraction is nothing other than a supersession (sublation) of existing ideas to better understand those ideas
that remain” (Kant, 1968a, p. 803, our transliteration). As Edmund Husserl (1976) argues in his text “The origin of
geometry,” although the ideal (more abstract) sense comes later than earlier results, “the earlier sense gives something
of its validity to the later one, indeed becomes part of it to a certain extent” (p. 373). In this sense, abstraction also
is a form of negative attention. It is in the dialectical negation of attention that new understanding emerges; without
this opposition or negation, Kant argued, no work would be required and the difference between existing and new
ideas could not be noticed. He further cautions us to distinguish pure ideas, which abstract from everything sensible
(“abstrahit ab omni sensitivo” [Kant, 1968c, p. 34]), and concepts given only empirically, which are to be denoted by
the term abstract (“qui autem empirice tantum dantur conceptus, abstractos nominare” [p. 34]).3 Empirical concepts,
such as color in the given example, are abstract; pure ideas (“Ideas puras” [p. 34]), however, cannot be abstracted from
the sensual world. For Kant, the issue is not whether concepts are abstract—“for every concept is an abstract concept”
(Kant, 1986d, p. 530)—but how concepts are used: concretely or abstractly. He does not privilege one over the other
because both are useful: “Through very abstract concepts we cognize a little in many things; by means of very concrete
concepts we cognize a lot in a few things—what we therefore gain on the one side we loose again on the other” (p.
531).
In his analyses how forms of thought could lead to knowledge and truth, Kant came to the conclusion that there had
to be knowledge that predated all learning. Time and space, for example, are not the result of sensual experience but
rather are presupposed by the senses. Thus, whether some sensual experiences are simultaneous or follow requires a
concept of time rather than the sequence leading to it. In particular, Kant postulated the cognitive capacity of critical
reflection prior cognition. Kant furthermore postulated the transcendental nature of self-consciousness, which means
that certain categories of thought only existed in the individual human mind and were not related to the material world.
This leads to a disjunction of concepts from the world of experience. Rightly so, this position, which constitutes the
foundation of Piaget’s work and (radical) constructivism, was recognized as constituting “subjective idealism” (Hegel,
1975). The Kantian approach remains insufficient to theorize processes of scientific activity such as interpreting graphs,
the one our research participant Eddie conducted (Roth & Hwang, 2006). Because graphs express something about
the experienced world and therefore are of the order of empirical concepts rather than pure ideas, we need to begin by
articulating the relationship of the subject of consciousness (knowing) and its object of consciousness (knowing). In

3 Kant wrote both in Latin and German; the German Werkausgabe (W. Weischedel, ed.) makes available the Latin texts with translations that
preserve sentence constructions and connections between ideas that other translations most often do not according to the editor. Because the first
author has studied both languages and is fluent in the second, we base our reading of Kant on both language versions to provide the best possible
English rendering of the sense of the cited and referenced passages.
338 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344

proposing dialectics, Hegel intended to overcome subjective idealism and its presuppositions of a priori categories by
proposing a theory of the development of the individual and collective consciousness.

3. Subject, object, and cognitive development in materialist dialectics

Dialectics constitutes an alternative to the dominant approach that separated body and mind, individual and collec-
tive, and knowing from the things known. Already practiced by Plato, the dialectical approach to knowing, learning, and
development was fully developed by G.W.F. Hegel; his work influenced, via the writings of Karl Marx, (educational)
psychology generally (e.g., Holzkamp, 1993; Leont’ev, 1978; Vygotsky, 1986) and the psychology of mathematics
(education) specifically (e.g., Davydov, 1990; Lave, 1988). Concerning mathematical knowing and learning, The Mas-
tery of Reason (Walkerdine, 1988) took the Kantian and Piagetian positions to task, arguing instead that mathematical
(thinking) practices cannot be dissociated from other social and material practices. The same author suggests that there
are fundamental problems with the Piagetian idea that the “physical world is governed by logicomathematical laws,
which came to form the basis of children’s development of rationality through their action on and adaptation of their
mental structures to the structures describing the physical world” (Walkerdine, 1997, p. 59). This position forces us to
rethink the relation of subject and the world, which changes both the nature of the subject and the nature of subjectivity
(e.g., Henriques, Hollway, Urwin, Venn, & Walkerdine, 1984).
Present day dialectical approaches to cultural (cognitive) practices begin analyzing and theorizing data with the
presupposition of individual and collective agency; agency, however, cannot be understood apart from structure, which
includes both structures of mind (schema) and structures in the material and social world (e.g., Bourdieu, 1980; Sewell,
1992). Agency and structure are dialectically related—which means that they are both different (non-identical) and
the same, that is, expressions of the same superordinate unit. All present-day dialectical conceptions of culture and
cognitive practices can be traced to Phenomenology of Mind/Spirit (Hegel, 1807/2003, 1977).4 In this study of how the
consciousness mind develops in forming abstract concepts—to the point that it not only becomes conscious of itself
but also comes to be able to theorize itself—Hegel showed that to overcome the problems in the Kantian approach, the
subject of conscious activity and its object of consciousness have to stand in a dialectical relation, each presupposing the
other, each constituting the negative identity to the other, and both pertaining to an integrative unit. From this relation
also springs development, because “consciousness itself is the absolute dialectical unrest, this medley of sensuous
and intellectual representations whose differences coincide, and whose identity is equally again dissolved” (Hegel,
1977, p. 124 [¶205], original emphasis). The integrative unit is practical activity, and it is in practical activity that the
subject both expresses its knowledgeability and changes it (Lave, 1993). The relationship between subject of activity
and object of activity is important to Marx’s (1973) reformulation of dialectics, because mind concretizes itself in the
results of practical activity and the concrete world comes to be reflected in mind.
Two points are important to retain. First, central to dialectical psychology in the Vygotskian lineage is the notion of
the twofold appearance of the object—on the first plane, the object is a real, often physical-material entity and therefore
distinct from the acting self-consciousness; and on the second plane, it is an aspect of the conscious mind (Leont’ev,
1978).5 The two planes are not independent but, in concrete activity, presuppose each other. Most scholars however
appear to forget that the subject, too, appears twice—as the body of an individual and as self-consciousness of the
mind (Hegel, 1977). Second, the subject is not identical with the person because, given that the object has material and
mental aspects, the person really is “spread” over and across the material subject and material object; or rather, both
subject and object constitute aspects of the person.
The stated relationship between subject and object is central to the development of the conscious mind (Lektorsky,
1980). It is in and as object of its activity that the conscious mind comes to see and become conscious of itself. That
is, mind becomes conscious of the world and of itself through something other than itself. More so, by becoming
conscious of the world, the mind not only learns about the world (object) but also about itself (Hegel, 1977); it is only

4 There are different renderings of Hegel’s German title Phänomenologie des Geistes, one rendering Geist as “spirit” (Hegel, 1977) and the other
as “mind” (Hegel, 2003). Although we find the latter rendering of Geist more appropriate, we actually prefer the former translation of the book as a
whole. In the preparation of this article, we used both translations as well as the German original (http://www.gwfhegel.org/PhenText/compare.html).
5 In the German philosophical literature, the term Gegenstand (object) denotes both material and mental entities, whereas Objekt (object) pertains

to material objects only. Hegel, Marx, and the psychologists following in their path use the term Gegenstand, thereby always connoting the double
nature of the object.
 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344 339

by objectivizing itself in the (material) results of its actions that a subject can know who it is (Ricœur, 1992). Such
learning and development, however, does not just occur, on its own, but requires societally mediated, object-oriented
and object-motivated, practical activity (Marx, 1976) because “society [i]s the first event of being” (Levinas, 1996, p.
23). That is, doing something—like Piaget’s child mathematician moving about pebbles—in itself is insufficient for
the development of anything like mathematical practices; rather, practical activity means doing something that has a
societally motivated object (purpose)—e.g., contributing to research, working at a checkout till, or calculating year-end
grades—and interacting with others. The famous experiments that the Soviet psychologist Alexander Meshcheryakov
conducted showed that deaf-blind children—who were more or less vegetating and, despite their ability to move
about and sense the material world, did not develop anything like human capacities—began to develop and eventually
even write books about their experience when others mediated their interactions with the world in purposeful ways
(Bakhurst & Padden, 1991). These experiments demonstrated that “the human being, as the subject of consciousness,
goal-directed creative activity is formed in intercourse with other people, the modes of which develop historically
and the means of which preserve in themselves the universal (social) determinates of all the objects of this activity”
(Mikhailov, 1980, pp. 259–260, our emphases).
Applied to the case study of scientists interpreting graph, this has at least two implications. First, the graph inter-
pretations do not stand on their own. The think-aloud session has to be understood as a form of realizing social science
research, a legitimate form of societal activity. Eddie and the other scientists do not just talk about the graph, they
also and foremost realize a think-aloud protocol, an instance of a societal activity. Second, interpretation has to be
understood as a continual movement of consciousness that goes from subject to object and from object to subject. This
aspect stands out in our case study featuring Eddie’s interpretation of one graph (Roth & Hwang, 2006).

4. Aporias of abstraction

In dialectical logic, the abstract, alternatively denoted as the universal or general, is thought and theorized differently
than in classical logic (Il’enkov, 1977). In classical logic, the abstract is thought in terms of the common properties
of a set of entities (objects, actions), which are then encapsulated in the superordinate concept (category). The classic
studies of concept identification involving drawings—rectangular boxes that contain various objects—that varied on a
number of dimensions (Bruner, Goodnow, & Austin, 1956) were based on this notion. Returning to our simple example,
in Bruner et al.’s experiments, children were expected to extract the concept of color given instances and non-instances
such as the following:

(3)
(4)
A key problem (aporia) with generalizing from common properties in this way is exemplified in the following
example of family resemblance (Wittgenstein, 1958). Given Smith-A and Smith-B have the properties a, b, and c
in common; Smith-B shares with Smith-C the attributes b, c, and d; Smith-D and Smith-A share property a; and
Smith-E and Smith-A have no attribute in common but their name. There is therefore no single attribute that all
Smiths share, though any sub-sample of Smiths may have none, one, or more common attributes. Yet, because it is
not difficult to construct properties for any participant sample, this situation poses fundamental epistemological and
empirical problems to the psychology of learning (Holzkamp, 1991). That is, all the Smiths might be related in a more
fundamental way: If Smith-A were a parent of the others, all Smiths thereby would be related even if they had no
property in common. Being historically and genetically the precursor of all others, Smith-A would then constitute the
universal, abstract, or general (Il’enkov, 1977).
In this approach, two important points need to be retained. First, the abstract, universal, and general (Smith-A) is
as concrete as the singular, concrete, and particular (Smith-B, -C, -D, -E). Any member of a further generation would
be equally singular, concrete, and particular: Each member of the filial generation merely realizes in concrete form
possibilities already present in the parent. Development therefore embodies a contradiction: It is a movement from
abstract to concrete but remains concrete throughout the change: “The method of rising from the abstract to the concrete
is . . . the way in which thought appropriates the concrete, reproduces it as the concrete in the mind” (Marx, 1973, p.
101). This statement is consistent with our experience described in the opening part of the companion text (Roth &
Hwang, 2006), where vectors first were merely concrete images and words on a page in a high school mathematics book
340 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344

(abstract), and subsequently were concrete, almost palpable entities in the hands and minds of a mature physicist and
statistician. Marx’s original framing therefore is opposite to the recently published interpretation some mathematics
educators offered (Hershkowitz et al., 2001). Applied to learning and development, early forms of consciousness
constitute historical precursors and contain possibilities that can be realized in different ways; this historical pattern
exists at both the individual and the collective level and is a condition for the objectification of geometrical knowledge
(Husserl, 1976). Because of the different ways in which possibilities can be realized, identical twins develop different
forms of understanding and form different identities despite identical starting points at birth. Their starting point
constitutes abstract, universal, and general possibilities that are concretely realized in different ways.
Second, being concerned with the development of mind, dialectics allows consciousness to overcome its own logical
contradictions in a new understanding that includes the previous understanding and also its contradictions in an abstract
and abstracting concept. To elaborate this point, we return to our example of concept learning. To learn concepts such as
“triangle” and “square,” the children in our second-grade classroom are exposed to sets of shapes. Among the different
possible ways of putting the shapes together, the teacher wants students to arrive at the following grouping:

(5)
However, the children initially tend to group in this way:

(6)
That is, children sort (categorize) the object by color or size. In the lessons we observed, the teacher actively discouraged
color and size: “You are not allowed to group by color or size.” The children therefore are asked to categorize according
to a category that they do not yet have developed. As Piaget (1970) noted, children in the early grades can carry out
the classification of these shapes without however understanding the class inclusion. To do the classification in (5) as
class inclusion presupposes the concepts triangle or square (i.e., one or both). The contradiction is that the task was
designed for them to arrive at the concept as the end result. Grasping (5) as a solution to the problem that the teacher
posed here cannot be the subsumption of the concrete triangles and squares; and even induction and abduction require
frames of reference to hypothesize the concept (category, rule) that when applied to the concrete case (here set (1))
leads to the desired result (here abstraction in the two sets of (5)) (Eco, 1984). “Because ‘the only way to grasp an idea
is to produce it”’ (Merleau-Ponty, 2002, p. 7) grasping the concept here is one movement that simultaneously goes
from concrete (set (1)) to abstract (sets (5)) and vice versa.
Thus, a child initially encounters entities in the world as sense-certainty, which it denotes as “this” or “that,” or
alternatively, especially prior to its ability to speak, to which it points. Here, any entity denoted as this, that, or through
pointing constitutes “mediated simplicity, or a universal” (Hegel, 1977, p. 61 [¶98], original emphasis). The entity
is both concrete and universal; but because it is a mediated universal, it necessarily has “to be the thing with many
properties” (p. 67 [¶112]). As the child learns to denote different entities of its experience as “dog,” “cat,” and “hamster,”
its understanding increasingly particularizes; at the same time, this particularization establishes the experiential and
conceptual foundation for, and simultaneous emergence of, the concept “mammal.” The process of abstraction does not
occur at the moment when the new concept (word) mammal comes to classify a range of experiences with animals, but
these experiences themselves—as dialectical philosophical discussions of the relationship of “origin” and “beginning”
from Plato to the present day show (e.g., Derrida, 1981, esp. pp. 75–84)—are the necessary origin and beginning of
the process and therefore anything like the generalizations that Kant and Piaget described.
The double movement appears to be foreshadowed by Kant, who viewed in a mathematical concept, such as the
concept of a triangle he used, the result of a construction on the basis of an a priori, non-empirical intuition (Kant,
1968b) that results from “introducing into space a particular set of limits or boundaries such that a three-sided figure in
one dimension results” (van Kirk, 1986, p. 141). The construction of an empirical (sensual) concept requires perceiving
that which is unique and unchanging about the figures in (1) so that it can be left behind as the non-unique (that which
they share) in the comparison. Upon reflection in one consciousness how different presentations are possible, all those
aspects in which different things differ can be dropped (Kant, 1986d), so that the described process of abstraction
therefore requires cognition of the singular.6 This implies, therefore, that the children’s grouping of the empirical

6 Constructivists recognize the former process but are unable to deal with, and, despite their Kantian heritage, have neglected attention to, the
radical singularity of sensual experience, which nevertheless is at the heart of empirical concepts.
 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344 341

triangles under the mathematical concept of the triangle requires two simultaneous movements: one through an ideal
plane, the other through an empirical (sensual) plane.

5. Developmental movement: a simultaneous double ascension from concrete to abstract and abstract to
concrete

In this article, we investigate the apparent contradiction between statements about learning and development as two
different, even opposing movements: development constitutes either an ascension from concrete to abstract (classical
version in the Kantian lineage) or an ascension from abstract to concrete (in the Hegelian lineage of dialectical position).
While the theoretical background of most research on mathematics and mathematics education falls into the former,
one recent study has touched on the contradiction by proposing to replace the former movement as the ascension from
a less to a more developed form of the abstract as concrete (Hershkowitz et al., 2001). Other mathematics educators
have followed in this same path (e.g., Monaghan & Ozmantar, 2006). However, in (materialist) dialectical logic, the
contradiction arises from the fact that the two ascensions are negating aspects of the subject|object movement, each
of which is realized only through its opposite in unity (Il’enkov, 1982); differences of degree constitute incomplete
negations and therefore cannot explain conscious thought or its development.
From a dialectical perspective, it is readily accepted that a theoretician mentally reconstructs the world in terms of
concepts developed by means of analysis. This reconstruction by definition is less full than the reality to which it refers
and therefore constitutes an abstraction that is the result of a generalization and analysis of immediate sense data. “In
this sense, and in this sense only, it is the product of the reduction of the concrete in reality to its abstract abridged
expression in consciousness” (Il’enkov, 1982, p. 136). It is a self-evident condition for any theoretical consciousness
to exist. However, the synthesis of separate abstractions yields the concrete in thought, which is formed in an ascent
from the abstract (in thought) to the concrete (in thought). In this way, “the ascent from the concrete to the abstract
and the ascent from the abstract to the concrete, are two mutually assuming forms of theoretical assimilation of the
world, of abstract thinking. Each of them is realized only through its opposite and in unity with it” (p. 137). The two
movements have to presuppose each other because otherwise (a) the ascent from the concrete to the abstract would be
“purely scholastic” thinking and (b) the reduction of the concrete to the abstract would yield a “disjoint heap of meager
abstractions” (p. 138).
We therefore propose to rethink learning and development in terms of a movement that is simultaneously from
abstract to concrete and from concrete to abstract. In dialectical fashion, the two contradictory aspects are but moments
of the same movement. Each moment is but a one-sided expression of the whole (developmental movement). That is,
we are not facing two concurrent movements one from the abstract to the concrete and the other from the concrete to
the abstract. Rather, we have but one developmental movement that appears to be in one direction or its opposite. The
distinction between the two apparent movements are superceded (“sublated”) in practical action—both in our case
study and in our classification example, learning coincides with the trajectory of interpretive action. In acting on an
object (graphs, set (1)), the conscious mind comes to realize and confront itself; in turning upon itself, consciousness
produces further realizations of itself that it compares with the object. In the process, it learns and develops. Because of
this dialectic relation of subject and object (i.e., subject|object as dialectic) of which the notion is found in the concept
of the double appearance of object and subject—one in consciousness and the other as material, any moment of human
activity is both abstract (reflection of reality in consciousness) and concrete (material reality, including our physical
bodies). The subject|object dialectic therefore lies at the origin of the double development from abstract to concrete
and concrete to abstract.
In our companion article, the activity the scientists such as Eddie are involved in is social science research; the partic-
ular task that they have taken up is the interpretation of graphs that we, the researchers, had culled from undergraduate
textbooks and presented them with. Two, mutually exclusive processes sublated in interpretation constitute the double
ascension of abstract and concrete in activity: practical understanding and explaining (Ricœur, 1991). These two pro-
cesses, however, are but one-sided expressions of the same activity that sublates both (Fig. 1). Practical understanding
derives from generally nonthematic and unthematized sensual experience of the world; explaining means employing
concepts to articulate and make salient structures in practical understanding. Thus, practical understanding precedes,
accompanies, and concludes the process of explaining; but explaining articulates and develops practical understand-
ing. Explaining moves from the abstract (concepts) to the concrete (e.g., the text to be interpreted), whereas practical
understanding moves from concrete to abstract. Whereas “concrete” denotes not only the materiality of graph but also
342 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344

Fig. 1. Practical understanding and explaining are two contradictory expressions of the movement constituted in and by the process of interpreting
a graph.

the lived experience of a graph reader that enables to begin interpreting, “abstract” denotes the salient aspect of lived
experience immediately brought about by directing toward the material graph and associated experiential concepts
articulated thereby. We visually present this view of interpretation as the contradictory movement from concrete to
abstract and abstract to concrete in Fig. 1, where, consistent with the cultural-historical activity-theoretic conception
of the subject and object, the two constitute a dialectic unit that appears twice: in material and “ideal” form.
It is in and through the movement of interpretation that the scientists in our studies on graphing, such as Eddie,
came to grasp the sense of the graphs. This grasp (conception) does not constitute the subsumption of a particular,
this graph, under a generality, a graph (graphs in general). Rather, grasp “is precisely the movement that negates the
general as well as the particular (movement that therefore also negates abstract relation), in order to affirm what alone
affirms itself in itself and for itself: the concrete singular, here and now, the existent as such” (Nancy, 2002, p. 66).
Everything that can be grasped as part of the familiar world surrounding us, whether known or unknown, “is an object
of experience, formally an A and this A” (Husserl, 1945, p. 333, our emphasis), that is, both a general and a singular.
Grasping, therefore, is the grasping of the singular in its singularity such that the “judgment determines at the same time
the rule and the case, by grasping the situation in its singularity” (Ricœur, 1992, p. 175). And grasping the singular in its
singularity means to grasp what is unique and unchangeable about it, and therefore, how it relates to the unchangeable
aspects of other singular situations.

6. Coda

In the final paragraph of the previous section, we arrive at an answer to the question posed in the title of this article:
learning neither goes from concrete to abstract nor from abstract to concrete. Learning constitutes a movement that
negates the abstract (general) and concrete (particular, singular) and opens up a channel for sense, the double passage
of the ideal (abstract) in the sensible (concrete) and the sensible in the ideal. The main point to retain is that the double
movement is not a result of shifting the meaning of the abstract and concrete to arrive at two different movements but
that the single movement is different within and from itself. Much of recent continental philosophy has been concerned
with establishing an ontology of difference, defined in and for itself rather than in terms of deviation from sameness
and self-identity (e.g., Deleuze, 1968/1994; Franck, 1981). In this respect, our approach of postulating learning as a
double ascension from concrete to abstract and vice versa is consistent with these recent philosophical efforts.
Here at the end, we deem it appropriate to make a reflexive comment. This double passage we describe—i.e., the
reverse ascension of abstract|concrete—pertains not only to the data we exhibited here, but also is characteristic of
the process by means of which we, the authors of this study, have arrived at our results. That is, we arrived at the
double ascension by being engaged in the detailed analysis of material aspects of human action, which constituted
our object, and in the process we articulated our own understanding of graphs, interpretation, and the development of
understanding.

References

Bakhurst, D., & Padden, C. (1991). The Meshcheryakov experiment: Soviet work on the education of blind-deaf children. Learning and Instruction,
1(3), 201–215.
 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344 343

Bourdieu, P. (1980). Le sens pratique. Paris: Les Éditions de Minuit. (Available in English as The logic of practice. Cambridge, UK: Polity Press,
1990)
Bruner, J., Goodnow, J., & Austin, G. S. (1956). A study of thinking. New York: Wiley.
Davydov, V. V. (1990). Types of generalization in instruction: Logical and psychological problems in the structuring of school curricula. Soviet
studies in mathematics education Reston, VA: National Council of Teachers of Mathematics.
Deleuze, G. (1994). Difference and repetition (P. Patton, Trans.). New York: Columbia University Press. (First published in 1968).
Derrida, J. (1981). Dissemination (B. Johnson, Trans.). Chicago: University of Chicago Press.
Dreyfus, T., & Gray, E. (2002). Research forum 1—Abstraction: Theories about the emergence of knowledge structures. In A. D. Cockburn & E.
Nardi (Eds.), Proceedings of the 26th Conference of International Group for the Psychology of Mathematics Education, Vol. I (pp. 113–138).
Eco, U. (1984). Semiotics and the philosophy of language. Bloomington: Indiana University Press.
Franck, D. (1981). Chair et corps: Sur la phénoménologie de Husserl. Paris: Les Éditions de Minuit.
Hegel, G. W. F. (1975). The encyclopedia of the philosophical sciences. Part I: Logic. New York: Oxford University Press.
Hegel, G. W. F. (1977). Phenomenology of spirit (A. V. Miller, Trans.). Oxford: Oxford University Press (German original published in 1807).
Hegel, G. W. F. (1988). Werke Band 2. Frankfurt: Suhrkamp.
Hegel, G. W. F. (2003). Phenomenology of mind (2nd rev. ed.) (J. B. Baillie, Trans.). Mineola, NY: Dover. (Originally published in English in 1910
and revised in 1931 [London: Macmillan]; German original published in 1807).
Henriques, J., Hollway, W., Urwin, C., Venn, C., & Walkerdine, V. (1984). Changing the subject: Psychology, social regulation, and subjectivity.
London: Methuen.
Hershkowitz, R., Schwarz, B. B., & Dreyfus, T. (2001). Abstraction in context: Epistemic actions. Journal for Research in Mathematics Education,
32, 195–222.
Holzkamp, K. (1991). Experience of self and scientific objectivity. In C. W. Tolman & W. Maiers (Eds.), Critical psychology: Contributions to an
historical science of the subject (pp. 65–80). Cambridge: Cambridge University Press.
Holzkamp, K. (1993). Lernen: Subjektwissenschaftliche Grundlegung. Frankfurt/M: Campus.
Husserl, E. (1945). Die Welt der lebendigen Gegenwart und die Konstitution der ausserleiblichen Umwelt. Philosophy and Phenomenological
Research, 6, 323–343.
Husserl, E. (1976). Husserliana VI: Die Krisis der europäischen Wissenschaften und die transzendentale Phänomenologie. In Eine Einleitung in die
phänomenologische Philosophie. The Hague: Kluwer Academic Publishers.
Il’enkov, E. (1977). Dialectical logic: Essays in its history and theory (H. Campbell Creighton, Trans.). Moscow: Progress.
Il’enkov, E. (1982). Dialectics of the abstract and the concrete in Marx’s Capital (S. Kuzyakov, Trans.). Moscow: Progress.
Kant, I. (1968a). Werke II: Vorkrische Schriften bis 1768 Frankfurt: Suhrkamp.
Kant, I. (1968b). Werke IV: Kritik der reinen Vernunft Frankfurt: Suhrkamp.
Kant, I. (1968c). Werke V: Schriften zur Metaphysik und Logik Frankfurt: Suhrkamp.
Kant, I. (1968d). Werke VI: Schriften zur Metaphysik und Logik Frankfurt: Suhrkamp.
Kant, I. (1968e). Werke XII: Schriften zur Anthropologie, Geschichtsphilosophie, Politik und Pädagogik. Frankfurt: Suhrkamp.
Lakoff, G. (1987). Women, fire, and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press.
Latour, B. (1993). La clef de Berlin et autres leçons d’un amateur de sciences [The key to Berlin and other lessons of a science lover]. Paris: Éditions
la Découverte.
Lave, J. (1988). Cognition in practice: Mind, mathematics and culture in everyday life. Cambridge: Cambridge University Press.
Lave, J. (1993). The practice of learning. In S. Chaiklin & J. Lave (Eds.), Understanding practice: Perspectives on activity and context (pp. 3–32).
Cambridge: Cambridge University Press.
Lektorsky, V. A. (1980). Subject, object, cognition. Moscow: Progress.
Leont’ev, A. N. (1978). Activity, consciousness and personality. Englewood Cliffs, NJ: Prentice Hall.
Levinas, E. (1996). Proper names. Stanford, CA: Stanford University Press.
Marx, K. (1973). Grundrisse. London: Pelicans Books.
Marx, K. (1976). Capital volume I. London: Penguin Books.
Merleau-Ponty, M. (2002). In L. Lawlor & B. Bergo (Eds.), Husserl at the limits of phenomenology: Including texts by Edmund Husserl. Evanston,
IL: Northwestern University Press.
Mikhailov, F. (1980). The riddle of self. Moscow: Progress.
Monaghan, J., & Ozmantar, M. F. (2006). Abstraction and consolidation. Educational Studies in Mathematics, 62, 233–258.
Nancy, J.-L. (2002). Hegel: The restlessness of the negative. Stanford: Stanford University Press.
Ohlsson, S., & Lehtinen, E. (1997). Abstraction and the acquisition of complex ideas. International Journal of Educational Research, 27, 37–48.
Piaget, J. (1970). Genetic epistemology. New York: W. W. Norton.
Quine, W. V. (1995). From stimulus to science. Cambridge, MA: Harvard University Press.
Ricœur, P. (1991). From text to action: Essays in hermeneutics, II. Evanston, IL: Northwestern University Press.
Ricœur, P. (1992). Oneself as another. Chicago: University of Chicago Press.
Roth, W.-M., & Bowen, G. M. (2003). When are graphs ten thousand words worth? An expert/expert study. Cognition and Instruction, 21, 429–473.
Roth, W.-M., & Hwang, S. W. (2006). On the relation of abstract and concrete in scientists’ graph interpretations: A case study. Journal of
Mathematical Behavior, 25(4), 318–333.
Russell, B. (1938). The principles of mathematics. New York: Norton.
Sewell, W. H. (1992). A theory of structure: Duality, agency and transformation. American Journal of Sociology, 98, 1–29.
Staub, F. C., & Stern, E. (1997). Abstract reasoning with mathematical constructs. International Journal of Educational Research, 27, 63–75.
Tsamir, P., & Dreyfus, T. (2002). Comparing infinite sets—a process of abstraction: The case of Ben. Journal of Mathematical Behavior, 21, 1–23.
344 W.-M. Roth, S. Hwang / Journal of Mathematical Behavior 25 (2006) 334–344

van Kirk, C. A. (1986). Synthesis, sensibility, and Kant’s philosophy of mathematics. In A. Fine & P. Machamer (Eds.), PSA: Proceedings of the
Philosophy of Science Association (Vol. 1) (pp. 135–144). East Lansing, MI: Philosophy of Science Association.
Vygotsky, L. S. (1986). Thought and language. Cambridge, MA: MIT Press.
Walkerdine, V. (1988). The mastery of reason. London: Routledge.
Walkerdine, V. (1997). Redefining the subject in situated cognition theory. In D. Kirshner & J. A. Whitson (Eds.), Situated cognition: Social,
semiotic, and psychological perspectives (pp. 57–70). Mahwah, NJ: Lawrence Erlbaum Associates.
Wilenski, U. (1991). Abstract meditations on the concrete and concrete implications for mathematics education. In I. Harel & S. Papert (Eds.),
Constructionism: Research reports and essays, 1985–1990 (pp. 193–203). Norwood, NJ: Ablex.
Wittgenstein, L. (1958). Philosophical investigations (3rd ed.). New York: Macmillan. (Reprinted in 1978).
Wittgenstein, L. (1978). Tractatus logico-philosophicus: Logisch philosophische Abhandlungen. Frankfurt: Suhrkamp. (Originally published in
1921 in Annalen der Naturphilosophie).