`Authentic science': Enculturation into the conceptual blind spots of
a discipline
Lansdowne Professor
Applied Cognitive Science
Faculty of Education, University of Victoria, Victoria, BC, V8W
3N4 Canada
Paper presented at the 1999 annual meeting of the American Educational
Research Association, Montréal, Québec.
Abstract
In this paper, I focus on some of the invisible ways in which students
of ecology are enculturated to the ideology of their discipline. The unreflected
matter of factness of the discursive and mathematical representations in
lectures and textbooks makes the world appear to be nomologically decomposible
(into variables) which have clear, mathematically fully determined relationships.
Introduction
During the latter part of the 1980s, a paradigm shift occurred in thinking
about thinking. Whereas the predominant cognitive models emphasized the
metaphor of mind as an information processor--with the associated notions
of short term and long term memory, procedural and declarative knowledge,
production systems that could be implemented in digital, linear computer
systems (e.g., Anderson, 1985)--anthropological and sociological studies
at work and in everyday settings began to emphasize knowing as embodied
practice (e.g., Latour & Woolgar, 1986; Lave, 1988). Following these
studies, educators called for reform of current curricula in ways that
would allow students to engage in the practices of mathematics of science
rather than in attempting to memorize procedural and declarative knowledge
(e.g., Brown, Collins, & Duguid, 1989; Roth, 1995). These educators
therefore suggest that classroom learning environments are constructed
such that they allow students to participate in practice, and which therefore
enculturate students into the practices characteristic of the field. Habitus,
a system of structured structuring dispositions generates perceptions,
and therefore the field of possible (material, discursive, etc.) patterned
actions, the practices characteristic of a field (Bourdieu, 1997).
Enculturation into the authentic practices of a field operates not
just on the mind, but on the social and material body by means of which
we are grounded in the world (Bourdieu, 1992, 1997). Our bodies, qua habitus,
with its history and embodied properties is formed by the social and material
world. Because our bodies have the property of being open to the world,
and are therefore exposed to the world, are susceptible to be conditioned
by the world, formed by the material and cultural conditions of existence
in which it is placed from the very beginning (a condition Heidegger 1977
called "thrownness"). That is, it is through our bodies exposed to the
social and material conditions of our existence that we are subjected to
a process of socialization of which individuation is itself a product,
the singularity of the "me" worked out in the social relationships. Enculturation
therefore forms the habitus, itself the generating mechanism of practices.
Habitus itself, however, resists to being thematized. Habitus is what drives
an expert scientist to take this action over that, to use this statistical
analysis over another, this set of nomological distinctions over a competing
one. Because it resists thematization, habitus can only be acquired by
participating in the field.
Given that what is to be communicated consists essentially of a modus
operandi, a mode of scientific production which presupposes a definite
mode of perception, a set of principles of vision and di-vision, there
is no way to acquire it other than to make people see it in practical operation
or to observe how thisscientific habitus. . . "reacts" in the face
of practical choices--a type of sampling, a questionnaire, a coding dilemma,
etc.--without necessarily explicating them in the form of formal precepts.
(Bourdieu, 1992, p. 222)
Thus, "authentic science," if it is to enculturate students, requires
them to operate under conditions that have, at least, some resemblance
with the conditions under which scientists operate. However, we migh want
to ask whether this is what we want to do in our schools. Unquestionably,
our physical and intellectual competencies are heterogeneous even within
our own fields: not every experimental physicist can do theoretical physics
well enough to publish in the field, or even publish in a another field
relatively close to her own; soccer, baseball, and football players all
have their specialty positions because each is associated with a particular
set of practices, and a particular habitus (often associated with the "instinct"
for shooting goals, finding the right research questions, etc.). However,
in educational practice, we appear to assume that "authentic practice,"
thinking like a scientist, mathematician, or historian is what we should
enculturate each and everyone into.
In recent years, science has come under increasing scrutiny because,
in many ways, it is not unlike traditional religions that requires its
followers to believe in a basic, unquestioned and normally unquestionable
set of presuppositions (e.g., Fuller, 1997), the reigning "ideology" of
the group (Ricoeur, 1991). However, even within the different subfields
of science, practitioners do not agree with all assumptions and question
the very pedagogy used to enculturate new members.
In this paper, I want to show how enculturate even in classrooms not
planned as authentic learning environments operates and produces a particular
habitus that in itself is often inconsistent with the field. Furthermore,
many students struggle with the inconsistencies between the scientific
ideology they are exposed to, and their own experiences.
Research Design
In this paper, I draw on our ongoing ethnographic work among middle school
students, college students, and ecologists. The data base includes early
200 hours of videotapes of field research, interviews, data analysis and
graph interpretation sessions, and lectures. We also conducted an intensive
text analysis of 6 high school and 4 university biology text books and
reviewed 5 annual issues of scientific journals in ecology (2,500 pages).
Assumptions
My basic assumption is that graphs (and their captions) constitute a form
of text (Eco 1976, Saint-Martin 1990); texts can themselves be regarded
within a semiotic framework where they function as the referents in relation
to the semantic depth analysis (sign) and the relevant interpretants (Ricoeur
1991). In my research team, our analysis begins with the act of reading,
and the (scientific) reader as an agent engaged in the analysis of texts.
This starting point is grounded in our daily experience as academics of
reading multi-modal (verbal, graphical, pictorial) texts individually.
Although language, text, and reading are shaped by the socio-cultural and
socio-historic context,[1]
individual readings of a text differ, turn up different text features and
relations, and contextualize these features in different, historically
and experientially contingent ways. This leads to considerable differences
in the `natural objects' that individual readers take the text to be about.
Our agent-centered framework considers reading as a semiotic process, enacted
by an individual with her own interpretive horizons, but who is always
and already embedded in a world shot through with meaning. This approach
also allows us to integrate the changing individual, so that our project
brings to the foreground the phenomenological, genetic, and hermeneutic
aspects of reading graphical displays.
Selecting the Domain
I chose as my domain of interest the representation practices in biology,
particularly ecology. There are three major reasons for choosing this domain:
(a) the levels of training in our team ranged from newcomer (high school
or college courses) to expert (MSc and research experience) providing an
analytical vantage point; (b) I had previously conducted research on the
representation practices of Grade 8 students engaged in ecological field
work (Roth, 1996; Roth & Bowen, 1993, 1994, 1995); and (c) my team
is conducting several studies on the representation practices in university
ecology courses and the training of Ph.D. students in the ecology of reptiles
(e.g., Roth & Bowen, 1998). I elaborate on the first point in the following
paragraphs.
Textbooks and Lectures
My team analyzed a series of biology textbooks commonly used in Canada
and the US. Six teachers from four districts indicated the same two textbooks
were used at the junior and senior levels; further inquiries revealed that
these are the only authorized biology textbooks in the province. We added
two textbooks from the authorized list of texts in a different province
that were used in a school where two authors had previously taught. We
complemented these four textbooks with two popular American textbooks from
the Biological Sciences Curriculum Studies series.
We also videotaped an entire second-year university ecology course,
lectures and seminars. The 13 week course had 45 students enrolled from
a variety of science and general studies backgrounds and each week consisted
of three fifty-minute lectures and three fifty-minute seminar; the latter
each being attended by approximately one-third of the students. The lectures
were presented by an ecology professor and a teaching assistant (doing
doctoral work in genetics/ecology) conducted the seminars. The two had
co-taught the course twice before and each year made refinements based
on their previous experience. In addition, we obtained about 25% of the
students' exams, and conducted several interviews and problem solving sessions
concerning graphs with 7 of the 45 students.
Scientist Participants
To understand how scientists read and come to read graphs, we recorded,
over a period of two years, university students at different levels of
training in science and scientists as they read graphs which we had culled
from a second-year university, entry level course in ecology. Two of these
graph-related tasks are featured in Figure 1. Our analysis of ecology textbooks
(Roth et al. 1997) showed that the type of graphs (distributions, isographs,
and graphical models) used in our research are very common in science.
These three types made for 120 graphs in a popular, 800-page ecology textbook
(Ricklefs 1990). We videotaped all 39 lectures and 36 seminars of the ecology
course; during the seminars, students frequently engaged in collective
interpretation of line graphs. On 7 occasions, we videotaped individuals
and small groups as they interpreted Cartesian graphs of our interest.
Ten preservice elementary teachers with an emphasis in science teaching,
who were in the process of completing a 5-year program, interpreted the
graphs in pairs. Four college science graduates who were presently teaching,
also interpreted the graphs. Finally, 16 practicing scientists with a minimum
of 5 years of independent research experience were recorded reading the
graphs we presented to them. In addition, scientists read graphs from their
own work.
Inquiry Process
All videotapes were transcribed in an ongoing manner so that the text was
available in written form during our analysis. We conducted analyses both
individually and collectively. Our analyses, grounded in semiotics and
hermeneutic phenomenology, are based on the assumption that reasoning is
observable in the form of socially-structured and embodied activity (Garfinkel,
1991). In our analyses, videotapes, transcripts, and artifacts produced
by the observed individuals are natural protocols of their efforts in making
sense of, and imposing structure on, their activities. These protocols
constituted our texts that we structured and elaborated in our analyses.
In the analysis of the graphs, I draw on the semiotics of multimodel
texts as a referent (Bastide, 1990; Lemke, 1998). I followed the advice
of hermeneutic phenomenology to produce alternate readings of the data
and thereby engage in a dialectic of understanding and explanation (Ricoeur,
1991).
Line Graphs
Line graphs are central to scientific practice because they can summarize
large data sets and suggest easily perceived topological structures in
data sets. Our analysis of textbooks had revealed that line graphs are
the predominant type of graphs in textbooks, which normally do not feature
plots of data set actually collected in research, or generated for didactic
purposes. For the purposes of this paper, I will draw on our research with
one type of line graph often used by theoretical ecologist as simple model
(Figure 1).
In
the derivation of a logistic model, we assume that, as N increased, birth
rates decline linearly and death rates increase linearly. Now, let's assume
that the birth rate follows a quadratic function (e.g., b = B0
+ (kb)N - (kc)N2), such that the birth
and death rates look like the figure. Such a function is biologically realistic
if, for example, individuals have trouble finding mates when they are at
very low density. Discuss the implication of the birth and death rates
in the figure, as regards conservation of such a species. Focus on the
birth and death rates at the two intersection points of the lines, and
on what happens to population sizes in the zones of population size below,
between, and above the intersection points.
Figure 1. Typical graph used in high school and university textbooks
and lectures. Task used in our research with university students, teachers,
and scientists.
Semiotic Analysis
This figure is typical of many that appear in high school textbooks: it
is marked by the absence of many resources that would assist in reading.
Scales and units on both axes, unit of analysis, the reference with respect
to which rate has been established, an explicit interpretant for the abscissa
label "N," etc. My analysis of 5 ecology journals showed that there was
not a single example of this type of graph. That is, interpretive resources
that are made available on a routine basis in the scientific literature
have been omitted here. On the other hand, the signs /birth rate/ and /death
rate/ which name the two lines also have cross modal implications in that
they refer to things that the students are familiar with in other more
familiar settings. For example, medical reports on the daily news frequently
tell us what the death rate due to one or another disease is, the death
rates of neonates according to different sectors of our own culture, or
in cross-national comparisons.
The two lines are "clean," there is no indication what actual data
might look like for which this model graph is a referent. Thus, because
of the functional relationship between "N," population density, and each
rate (linear for the death rate, quadratic for the birth rate) there are
point form intersections between the two line.
The standard reading of the graph goes as follows: at each intersection,
birth rate and death rate are equal so that the population should be at
an equilibrium. However, the left and right intersection (Figure 1) differ
in that one designates a stable equilibrium whereas the other an unstable
equilibrium. Let us begin with the left equilibrium. A little to the left
of the intersection, the death rate is larger than the birth rate so that
the population density declines, leading to an even larger difference in
favor of the death rate, and therefore, ultimately, to a crash of the population.
The opposite trend occurs a little to the right of the left intersection.
Because birth rate is larger, the population increases which leads to a
new population with birth rates larger than death rates, which continues
until birth and death rates are equal again at the right equilibrium. This
equilibrium is stable because, as soon as a population becomes larger (because
of some random event), the death rate is larger than the birth rate, and
the population is pushed back to the equilibrium.
Embodied in such readings, of course, are a number of assumptions.
First, if birth and death rates follow exact relationships as presented
here, then there are equilibrium points for fixed population sizes. Thus,
these models presuppose that there are in fact stable and unstable equilibrium
points rather than some other state. For example, Ilya Prigogine (1980)
suggested that a better way to approach science is in terms of non-equilibrium
systems that are open to fluctuations and never have stable points.
Second, the model is deterministic. That is, given a population size,
we know what the end state of the population dynamic will be. That is,
whereas common ecological wisdom suggests that populations oscillate (Figure
2.top)--though even this is questioned by some and there are only few predator-prey
systems that behave in an ordered cyclical way, many populations are non-cyclical
so that more recently, patterns are searched using time series analysis
and neural network techniques--in the present case, however, the population
dynamic is to a constant, steady final state (Figure 2.center). In any
event, given the criticism steady state and cyclical models receive--see
also next section--other models more dynamic and open to random variations
may be more appropriate (Figure 2.bottom). Such discussions are usually
absent from any ecology course.
Figure 2. Top. One common ecological paradigm suggests
that populations go through cycles. Such cycles can be modeled by a two-organism
predator-prey model. Center. Actual population dynamic predicted
by the graph in Figure 1. Bottom. Model of Figure 1 with random
variations of on birth and death rate.
Third, the model presumes that it is feasible to express a population
(dynamic) in terms of birth rates and death rates, irrespective of whether
and how the corresponding measures can be taken. From a measurement perspective,
for example, birth rates and death rates are not homogeneous throughout
the year and for all members of a population. Thus, rut and birthing usually
occur during a particular time of the year. The dynamics of birth and death
rates, and therefore of the population is driven by the contingencies of
the setting, climatic conditions, killing of the weaker (old, young) in
a population, the incidence of the birth date relative to the coming seasons
which support or interfere with the development of the young, etc. Some
may argue that birth rate and death rate can be accumulated over longer
periods such as a year. However, the very dynamics of the mathematical
model and the population dynamics it therefore has as a referent, depends
on the assumptions about birth and death rates and how they are accumulated.
Fourth, the model does not suggest that the implications for population
dynamics are different--and therefore to conservation or harvesting policies--if
birth and death rates indicated by the function are only best estimates
of some mean value of data that vary considerably. If one considers variation,
we need to consider birth and death rates as bands generated by the confidence
intervals around the curves. As Figure 3 shows, with a representation that
includes variation to some degree, we no longer "perceive" stable equilibrium
points but ranges of population size values. Ranges immediately
change our interpretation of the lower equilibrium point in the sense that
it is crucial not to lie inside this zone if one is considering to maintain
the population in both the conservation and harvesting paradigms.
Figure 3. Population graph including variability. There are no
longer equilibrium points, but critical zones.
Fifth, the very use of a population model in the form of a graph suggests
a particular epistemology. This epistemology is embodied in the isomorphism
of the form {Fundamental Structure <--> Mathematical Structure} which
is taken in the sciences--in keeping with the Greek tradition of seeking
and presupposing a mathematical order of the universe--as a fundamental
paradigm (Lynch, 1991).
Finally, the very use of smooth graphs enculturates students to an
expectation that natural phenomena are inherently mathematical. Thus, when
they face a set of data that do not fall onto a clear and simple increasing
or decreasing pattern, most students have difficulties interpreting the
data and propose that there are other things affecting the relationship
between two sets of measurements on the same unit of analysis. In a series
of experiments, we tested (a) Grade 8 students who had done their own research,
including the transformation of data into a variety of representation in
order to convince their own peers and (b) preservice teachers who were
in their fifth professional year of study and almost all of whom had previously
obtained a B.Sc. or M.Sc. degree (Roth, McGinn, & Bowen, 1998). Both
groups were given a task involving a map subdivided into sections, each
of which contained two sets of measures: light intensity and plant density.
The participants were asked to help the student who had generated the data
in deciding whether there is a relationship between the two measures. Statistical
comparisons revealed that there was a significantly higher proportion of
Grade 8 students (who solved the problem in pairs) using graphical and
statistical analysis methods than secondary preservice teachers. Having
classified responses into more abstract representations (graph, averages),
less abstract representations (ordered table, pattern map, list), and no
transformation (language-based), we detected a significant effect ([chi]2(2)
= 6.80, p < .05). There was a lower incidence of more abstract
representations among preservice teachers than among pairs of Grade 8 students.
Furthermore, there was a relationship between the type of analysis and
the type of claim respondents made. A logit analysis-with type of claim
(correlation, no correlation) as dependent variable and type of representation
(more abstract, less abstract, none) as independent variable-showed that
an equi-distribution model had to be rejected, [chi]2(3)
= 16.42, p < .001. Analyses based on statistical and graphical
methods generally suggested a positive correlation, whereas analyses based
on other methods generally ended in claims that there existed no relationship
between the two variables.
We can understand this example in terms of a habitus in wich relationships
between measures that objectify the world are mathematical rather than
trends that summarize stochastic data. These students were enculturated
such as to expect cycles and, at best, were willing to conceived of epicycles
("other variables") that mediated the kind of relationship they expected
to exist. When all is said and done, that is, when all variables are accounted
for, these students expected there to be no unexplained variation to remain.
Scientists' Appreciation
Although graphs such as that represented in Figure 1 is standard fare of
science textbooks from high school to university, not all scientists agree
with the "implicit" meanings they associate with these graphs. In our interviews
with the scientists, about half of the participants critiqued the population
graph for one reason or another, and in particular in respect to the kind
of ecological thinking students are enculturated to.
In some situations, the scientists evaluated the graphs at a more global
level and labeling the graphs as biologically unrealistic (didactic) cases
that had little relevance or meaning in "the real world of science." Thus,
based on specific surface features, these scientists argued that there
could be no meaningful referents in the real world which is the main concern
of their own work: the set of referents is an empty set. In this instance,
the lack of scales and unit of measurement, the lack of data points, the
simplistic topology of a curve, unconventional axis labels and other features,
the lack of stochasticity and so forth were global features which "told"
some individuals that the graphs were of little value to real science.
Figure 4. Graphical model to represent "essential resources"
as it was used in the second-year university ecology course.
For example, the population dynamics graph led some scientists to the
conclusion that it was unrealistic: it did not portray natural variations
in the rates (error) which might mislead interpreters (managers) to the
(false) conclusion that populations are save from extinction as long as
they are above the equilibrium, however close. Thus, one ecologist (successful
in terms of grants received and publications) suggested:
[SA:] You're never gonna find a data set that looks like this. This
is a theoretical model, it's based on, you know, nice mathematics and equations,
and it's the way we think the world probably works. But I don't know any
data set where you ever find this and you can ever point out there are
probably two steady states. It's just, in the real world it is a constant
fluctuation. There's two ways to approach the problem, one is, I mean,
you collect real data then look where the patterns are, and you might,
I mean, it's theoretically possible to create something like this from
real data, the other approach is to start at the other end, and just to
start with pure mathematics and conceptualize how these populations must
change, use equations to describe it, and then go on for data that might
fit this to try to validate your hypothesized model, so you can approach
it in two different ways but I've never seen in this case. I don't know
any example where this fits.
Others suggested that natural situations are much more complex as to
be modeled by a density dependent function alone. Thus, once a scientist
begins with the assumption that a particular graph cannot have a referent
in the world, their work of establishing possible referents already experienced
an interference.
In another example, four scientists explicitly referred to the sharp
corners in the "essential" resource graph (Figure 4) as biologically unrealistic.
One of these noted:
[MC:] I'm just puzzled by these sharp, I'm not, never, I'm not accustomed
to seeing the sharp, sharp bends. It's like a threshold or some kind. .
. In biology, we don't usually get that kind of a pattern, at least I'm
not accustomed to seeing a pattern like this.
In summary, then, many field ecologists disagreed with what they considered
to be the epistemology implicit in these graphical models. This epistemology,
in their view, interferes with the holistic understandings they themselves
associate with the discipline and which is captured in its name (eco-logy,
from Gr. oikos, house). Another aspect of forming the habitus of new genereations
of researchers lies in the way nature is decomposed into variables.
Nature of Variables
Traditional philosophy of science lore has it that identifying and controlling
variables are core "scientific process skills." The innovative curricula
of the 1960s and 70s (including SAPA [II] and SCIS) included special activities
in which children were asked to identify and control variables. However,
the ethnographic descriptions of scientific laboratories do not confirm
these as aspects of laboratory work--though these are recognizable features
of scientific articles. Rather, what is a suitable variable often emerges
as scientists become increasingly familiar with the domain of current interest.
The Canadian genticist David Suzuki (1989) described this as an inappropriate
image of how scientists do their work. Describing an important discovery
by his own laboratory, Suzuki suggested that they had started to look at
a protein defect (which might explain the temperature sensitivity of a
certain fly). Their work began with speculations and initial experiments
in the course of which many new questions became salient and themselves
researched. With each new questions, new and different ways of carving
up the world in terms of variables seem to emerge.
By emphasizing a propoer way to do an experiment and to write ut up,
we create a myth about how research is done. And we lose all the passion
that makes the scientific enterprise so worthwhile. I hope the new science
curricula don't make this mistake. (p. 192-3)
Working on the frontier of the known objectified world, they searched
in the dark. Even in traditional laboratory exercises students work "in
the dark," apparently caught in a double bind. To know whether what they
observed is what they should have observed they need to know that what
they have done was what they should have done. At the same time, to know
whether what they have done is what they should have done they need to
know that what they had observed is what they should have observed (Roth
et al., 1997).[2]
My own research in "open inquiry" science classrooms (e.g., Roth &
Bowen, 1994; Roth & Roychoudhury, 1993) and among field ecologists
(e.g., Bowen & Roth, 1998; Roth & Bowen, 1998) shows that what
are possible and viable ways of objectifying the world only arises with
experience in the particular domain or part of the natural world. Thus,
variables are not just identified and perceived as stable structures in
the world, imposing themselves on the observant scientist or science students.
Rather, variables emerge from the coincident fit of natural constraints,
manipulations of equipment and materials, and the (descriptive and theoretical)
discourses we use to objectify our experiences (Gooding, 1992; Pickering,
1995).
In contrast to the cited work, traditional science courses at the high
school and university levels alike suggest that nature has a stable true
set of joints along which it has to be cut in the very way they present
graphs, discourse, nature, etc. Thus, traditional teaching enculturates.
Let us then take a look at the following example.
Variables are Projections
A Case in Point: The Resource Lecture
As an example of the unquestioned manner in which variables in graphs were
presented as decontextualized facts rather than as the result of protracted
experiences and modification of practices until observation, discourse,
manipulations were consistent with each other. In the lecture, there was
little discussion about the nature of any variables and the purpose of
their selection. To the students in the ecology course the reductionism
embodied by such a treatment was invisible such that the variables were
expected to be accepted as real. One part of the resource lecture focused
on substitutable resources and was accompanied by a graph said to display
this kind of resource (Figure 5), the array of examples might well lead
students to understand that there was a considerable number of possible
resources to look at, but no opportunity to try to understand why those
specific variables were chosen for discussion over any other possible variables.
Figure 5. Graph said to represent "substitutable resources."
From a semiotic perspective, the graph elaborates "substitutable resources"
in its function as interpretant, and signifies (as referent) some situation
itself only available through situation descriptions.
The following excerpt from the lecture transcripts illustrates this
argument as the lecturer provides a literal reading of the graphs depicting
resource interaction:
[Professor:] Substitutable resources, they are substitutable, either
can replace others, they may not be equally good. So if you're a lion or
cheetah, or on a savanna--well if you need zebras or if you need gazelles
they are both pretty much as good. Similar to each other as resources [you
can] either live on, over a year, 30 gazelles or maybe 20 zebras. And another
example of this is when you're on an airplane and flight attendants come
down the aisles and say `Chicken or fish, chicken or fish, chicken or fish.'
Well they're substitutable in here, it may be better to say, `They may
not be equally bad.' This isocline shows substitutable resources. If you
have a little bit of R2 and quite a bit of R1 the
population will do about as well as if it has a little bit of R1
and quite a lot of R2. So you can substitute one for the other.
They may not be good but we can substitute.
This transcription is rather typical for the way in which students
experience the presentation of variables (i.e., types of resources) as
the joints along which the world has to be cut without any
context of why these particular cuts were made. The resource lecture presented
students with a considerable number of "examples" of the different outcomes
from the interaction between two nutritional resources. Yet, even in the
talk of resources that preceded it there was little mention of the variety
of resources necessary for organisms to survive. The various examples offered
represent a bewildering range of categories, units of measure, and outcomes
of effect (e.g., potassium, calcium, nitrogen, magnesium, pounds, kilograms,
growth, growth rate, maintenance of health, etc.). There existed few opportunities
for students to learn about the range of possible resources specific to
an organism and therefore they had little context within which to disclose
and develop their understanding of the relationship between two resources
when re-constructing the larger picture that constitutes understanding
a "realistic" situation.
In this presentation, variables are presented as a matter of fact;
there is little discussion about how and why any specific pair of resources
matter in the study. Our interviews with scientists showed that they usually
need such context to re-situate the results of any one comparison within
a broader understanding of ecology and resource use (Roth, Masciptra, &
Bowen, 1998). Choice of which resources would be used for study would be
unlikely to occur for arbitrary reasons and yet the choice of specific
variables for interaction studies is not a transparent process for newcomers
to the discipline such as the students in the course. The lack of descriptive
context is therefore particularly problematic for students in a class where
students, for the first time, learned about resources and their interactions
and, also for the first time, faced a two dimensional depiction of those
interactions. Students are listening to the talk about ecology as newcomers,
without the insight that field experience lends to the comments being made,
listening to lecture talk that is structured as if old-timers were telling
their autobiography.
Submitting to Discipline and Being Disciplined
Our analyses of students' discourse surrounding graphs showed that they
bought into the reductionist approach. For example, they spent considerable
time and effort trying to place the population graph (Figure 1) in an ecological
context that would reflexively elaborate graph and context (e.g., Roth
& Bowen, in press). However, their discussions also suggested that
they had insufficient experiential resources relative to ecology to aid
their interpretations (Bowen, Roth, & McGinn, in press). Thus, why
these variables rather than others might be an important aspect of a lecture
designed to help students to reflect in considering the biological "reality"
of the situation. In the present case, it matters what type of species
of organisms are involved, for motility allows the organism to re-cluster
into smaller geographic areas and thereby increase their densities to the
critical level needed to survice. Within ecological reasoning, the ability
to move and re-cluster is an important aspect of whether the situation
is biologically realistic or not. Yet, students did not have the
interpretive resources to address any of these other variables. Furthermore,
any attempt to bring in other variables was stopped by the teaching assistant
who emphasized that they should only focus on the variables presented in
the scenario.[3]
Students did not make sense of the reduced model in the context of
a broader picture of other related variables. In their interpretations
of the population graph, death rate was taken at face value as a clearly
defined variable without discussion of the density-independent factors
that can affect it-such as variations in predator populations, food resources,
and other variables which themselves vary without correlating with population
density. This is in contrast to some of the scientists' discussions in
which the specifics of the determination of the variable "death rate" were
considered to be important to the specifics of their interpretation of
the graph. Students did consider the problem ceteris paribus, as
one in which all other external factors remained constant and only changes
in the target population are relevant. Complexities of variables, such
as occur when the varied influences on death rate are considered, were
reduced to being unitary in effect. We found, in subsequent focus-group
interviews, that this difficulty with embedding the graphs' variables in
the larger context of potential variables and influences seemed to considerably
hamper students' developing a plausible interpretation of the graph.
Discussion
Enculturation
A central issue for Bourdieu are the ways in which we become enculturated.
Enculturation is coincident with the formation of a specific habitus, a
set of dispositions that structure perceptions of and actions toward the
world, but are themselves structured by the experience in the world. By
participating in biology as it is taught in high schools and at the university,
students come face to face with particular ways of structuring the world--as
embodied in the discourses, mathematical representations, images, and so
on which form their habitus itself responsible for generating the practices
in which they enact. In our observations, university students develop expectations
(generated by habitus) that the world is inherently decomposable along
a set of a priori and therefore true joints (variables) and inherently
mathematical so that, after everything is said and done, no or only minor
variations are left unexplained. It is the enculturation to such a view
of the world that most of the field ecologists objected while at the same
time not questioning their own sets of assumptions.
Enculturation, however, does not come as easy as I presented it. For
our habitus does not simply change when we enter a new (sub-) culture.
Rather, as Bourdieu outlines, not any old habitus will facilitate enculturation,
but the new member has to bring is not the habitus which is already tacitly
and explicitly required, but a habitus which is practically compatible,
or sufficiently close, and above all malleable and susceptible to being
changed into a habitus that is conform (Bourdieu, 1997). We may find here
one way of explaining why there are differences along traditional lines
of difference (gender, economic status, etc.) as to the participation rates
in science at the high school and even more pronounced at the university
level and in professional life.
Scientific Representation Embody Epistemologies
High school textbooks and scientific journal articles also differ in the
way they treat Cartesian graphs with data and those that feature models;
that is, there is a difference between data which are used to ascertain
or to induce some relationship and graphs which express a relationship
that was deduced from a theoretical model. In scientific publications,
there is a recognizable relationship between "real" data and the theory
that is expressed in continuous line graphs; behavior is reduced to a single
law or set of laws. The clean line of the smooth curve is given authority
by the fact that the individual measurements do not fall on the line, and
frequently are associated with bars that indicate errors of measurement.
Points suggest accuracy and statistical variation, whereas the smooth curve
approximating them holds out the hope for a simple relationship that can
be expressed in a mathematical equation. Smooth curves displayed alone
are looking for data points. Reviewers in the scientific community often
take graphs without actual data as "lazy attempts at demonstration" (Myers,
1990, p. 244). But such graphs are exactly what readers of high school
textbooks continuously face. This interaction between empirical data and
relationships of theoretical nature is no longer available to textbook
readers. Here, graphs are detached from empirical situations to which they
might relate. But textbook authors never make it clear that the featured
line graphs are used to express currently accepted models.
In the present study, I am concerned with biological texts. Biology
provides "ways of looking at nature" (Bjelic, 1992, p. 222). Biological
texts, therefore, both as research papers and textbooks, serve as demonstrations
in which readers are shown, and learn to see, the order reflexively specified
by the textual arrangement of words and graphics. The text, then, is a
worksite where knowledge is constructed not through the passage of information
from the book or article to the reader, but through the recovery of the
order encoded by the author in text and accompanying graphs. It is through
the noticeable work of reading that the knowledge of biological objects
is constructed. The convincing text engages readers in a process of authentication
so that readers recover what the biologist previously saw. Texts that achieve
this authentication, that is, texts that achieve the demonstrable fit between
graphs and text, display a compelling pedagogy in teaching readers something
about the world they inhabit.
Conclusion
While I do not have the space to engage in a discussion of a different
approach to teaching any science, I refer the reader to an article in which
Jacques Désautels and I recommend an epistemology as practice enacted
by high school students (Désautels & Roth, in press). In such
a science, students would not merely learn science (mathematics, history,
etc.) but as part of their activities of studying a domain also engage
in an epistemological critique which would assist them in re-contextualizing
and thereby relativizing all forms of knowing--both in terms of their affordances
as well as their shortcomings.
My research also raises a host of new questions. First, the questions
we answered about graphing practices are yet unanswered for other inscription
practices: How are [photographs, diagrams, tables, etc.] used in textbooks
and in scientific journals to inculcate a particular epistemology? and
What are the relationships between [photographs, diagrams, tables, etc.],
captions, and main text in to produce such inculcation? Second, empirical
work needs to be done to answer questions about students' appropriation
and developing competencies related to inscription practices in redesigned
classroom environments that are based on semiotic perspective which allows
students to enact a semiotic critiqie of representations and the epistemology
they support: How do students' semiotic competencies develop in learning
environments that emphasize semiotic analyses? and How do students' epistemological
practices change in learning environments where scientific concepts are
treated as heterogeneous assemblies?
Acknowledgments
This paper was made possible by a grant of the Social Science and Humanities
Research Council of Canada (410-93-1127). My thanks go to Michael Bowen
and Sylvie Boutonné who assisted in the collection and transcription
of the database on which I have drawn.
References
Anderson, J. R. (1985). Cognitive psychology and its implications.
San Francisco, CA: Freeman.
Bastide, F. (1990). The iconography of scientific texts: principles
of analysis. In M. Lynch & S. Woolgar (Eds.), Representation in
scientific practice (pp. 187-229). Cambridge, MA: MIT Press.
Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition
and the culture of learning. Educational Researcher, 18(1), 32-42.
Bourdieu, P. (1992). The practice of reflexive sociology (The Paris
workshop). In P. Bourdieu & L. J. D. Wacquant, An invitation to
reflexive sociology. Chicago, IL: The University of Chicago Press.
Bourdieu, P. (1997). Méditations pascaliennes [Pascalian
meditations]. Paris: Seuil.
Bowen, G. M., & Roth, W.-M. (1998, April). Isolation of variables
and enculturation to a reductionist epistemology during ecology lectures.
Paper presented at the annual conference of the American Educational Research
Association, San Diego, CA.
Bowen, G. M., Roth, W.-M., & McGinn, M. K. (in press). Interpretations
of graphs by university biology students and practicing scientists: towards
a social practice view of scientific re-presentation practices. Journal
of Research in Science Teaching.
Bjelic, D. I. (1992). The praxiological validity of natural scientific
practices as a criterion for identifying their unique social-object character:
The case of the 'authentication' of Goethe's morphological theorem. Qualitative
Sociology, 15, 221-245.
Désautels, J., & Roth, W.-M. (in press). Demystifying epistemology.
Cybernetics & Human Knowing.
Fuller, S. (1997). Science. Buckingham, England: Open University
Press.
Garfinkel, H. (1991). Respecification: evidence for locally produced
naturally accountable phenomena of order*, logic, reason, meaning, method,
etc. in an as of the essential haecceity of immortal ordinary society,
(I)--an announcement of studies. In G. Button (Ed.), Ethnomethodology
and the human sciences (pp. 10-19). Cambridge: Cambridge University
Press.
Gooding, D. (1992). Putting agency back into experiment. In A. Pickering
(Ed.), Science as practice and culture (pp. 65-112). Chicago, IL:
The University of Chicago Press.
Heidegger, M. (1977). Sein und zeit [Being and time]. Tü-bingen,
Germany: Max Niemeyer.
Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: The
MIT Press.
Latour, B., & Woolgar, S. (1986). Laboratory life: The social
construction of scientific facts. Princeton, NJ: Princeton University
Press.
Lave, J. (1988). Cognition in practice: Mind, mathematics and culture
in everyday life. Cambridge: Cambridge University Press.
Lee, S., Roth, W.-M., & Bowen, G. M. (1999, March). Learning
at the interface: Knowledge transfer and transformation between heterogenous
and overlapping knowledge-building communities. Paper presented at
the annual conference of the National Association of Research in Science
Teaching, Boston, Mass.
Lemke, J. L. (1998). Multiplying meaning: Visual and verbal semiotics
in scientific text. In J. R. Martin & R. Veel (Eds.), Reading science
(pp. 87-113). London: Routledge.
Myers, G. (1990). Every picture tells a story: Illustrations in E.
O. Wilson's Sociobiology. In M. Lynch & S. Woolgar (Eds.), Representation
in scientific practice (pp. 231-265). Cambridge, MA: MIT Press.
Pickering, A. (1995). The mangle of practice: Time, agency, &
science. Chicago, IL: University of Chicago.
Prigogine, I. (1979). Vom sein zum werden [From being to becoming].
München: Piper.
Ricklefs, R. E. (1990). Ecology (3rd ed.). New York: Freeman.
Ricoeur, P. (1991). From text to action: Essays in hermeneutics,
II. Evanston, IL: Northwestern University Press.
Roth, W.-M. (1995). Authentic school science: Knowing and learning
in open-inquiry laboratories. Dordrecht, Netherlands: Kluwer Academic
Publishing.
Roth, W.-M. (1996). Where is the context in contextual word problems?:
Mathematical practices and products in Grade 8 students' answers to story
problems. Cognition and Instruction, 14, 487-527.
Roth, W.-M. (1998). Designing communities. Dordrecht, Netherlands:
Kluwer Academic Publishing.
Roth, W.-M., & Bowen, G. M. (1993). An investigation of problem
solving in the context of a Grade 8 open-inquiry science program. The
Journal of the Learning Sciences, 3, 165-204.
Roth, W.-M., & Bowen, G. M. (1994). Mathematization of experience
in a Grade 8 open-inquiry environment: An introduction to the representational
practices of science. Journal of Research in Science Teaching, 31,
293-318.
Roth, W.-M., & Bowen, G. M. (1995). Knowing and interacting: A
study of culture, practices, and resources in a Grade 8 open-inquiry science
classroom guided by a cognitive apprenticeship metaphor. Cognition and
Instruction, 13, 73-128.
Roth, W.-M., & Bowen, G. M. (1998, October). Perceptual topology
of and mathematization in ecology fieldwork. Paper presented at the
annual meeting of the Society for Social Studies of Science, Halifax, N.S.
Roth, W.-M. , & Bowen, G. M. (in press). Complexities of graphical
representations during lectures: A phenomenological approach. Learning
and Instruction.
Roth, W.-M., Masciotra, D., & Bowen, G. M. (1998, October). From
thing to sign and `natural object': Toward a genetic phenomenology of graph
interpretation. Paper presented at the annual meeting of the Society
for Social Studies of Science, Halifax, N.S.
Roth, W.-M., McGinn, M. K., & Bowen, G. M. (1998). How prepared
are preservice teachers to teach scientific inquiry? Levels of performance
in scientific representation practices. Journal of Science Teacher Education,
9, 25-48.
Roth, W.-M., McRobbie, C., Lucas, K. B., & Boutonné, S.
(1997). The local production of order in traditional science laboratories:
A phenomenological analysis. Learning and Instruction, 7, 107-136.
Roth, W.-M., & Roychoudhury, A. (1993). The development of science
process skills in authentic contexts. Journal of Research in Science
Teaching, 30, 127-152.
Saxe, G. B. (1991). Culture and cognitive development: Studies in
mathematical understanding. Hillsdale, NJ: Lawrence Erlbaum Associates.
Scribner, S. (1986). Thinking in action: some characteristics of practical
thought. In R. J. Sternberg & R. K. Wagner (Eds.), Practical intelligence:
Nature and origins of competence in the everyday world (pp. 13-30).
Cambridge: Cambridge University Press.
Suzuki, D. (1989). Inventing the future: Reflections on science,
technology, and nature. Toronto: Stoddart.
Winograd, T. (Ed.). (1996). Bringing design to software. New
York, NY: ACM Press.
[1]
Derrida (1998) insists on the inherently social nature of all language:
"We only ever speak one language--and, since it returns to the other, it
exists asymmetrically, always for the other, from the other, kept
by the other. Coming from the other, remaining with the other, and returning
to the other" (p. 40). It is therefore not suprising that Livingston (1995)
suggests that reading is a social process known through its achievement
in the individual act of reading.
[2]
It may help readers to think about the following vernacular example. Novice
cooks find themselves in such double binds. If they are fortunate, their
cookbook provides a picture of the final result. Although they follow the
steps of the recipe, they cannot know whether what they have done is what
they should have done. However, many cookbooks have pictures which serve
as check points to exit the double bind. For, if their own final result
is dissimlar to that displayed, they know that what they have done was
not what they should have done.
[3]
Similar comments ("focus just on the information given") were provided
by scientists to grade 8 students in a study on representation practices
(Roth 1996). Such comments appear to be typical though implicit for individuals
who have been enculturated into reductionist worldview when they attempt
to enculturate newcomers into the same worldview.