Tag Archives: heterogeneous_construction

Three angles from which to view the practice of researchers II

My book, Unruly Complexity: Ecology, Interpretation, Engagement (Taylor 2005), considers three angles—like facets of a crystal—from which to view the practice of researchers (as introduced in the previous post).
Continue reading

Mapping: Can scientists become interpreters of science and bring the interpretations to bear on their science?

According to the perspective of heterogeneous construction, scientists mobilize a diversity of resources and, in so doing, engage with a range of social agents.  Similarly, when interpreters of science delimit the relevant resources and agents, they also mobilize resources and engage with diverse social agents (Taylor 2005, Chapter 5, section A). Interpreters of science who recognize this might then reflect explicitly on the practical conditions that enable them to build and gain support for their interpretations. Applying the same interpretive framework to one’s own research should enhance the plausibility of their reconstructions of the work of scientists.
There might be more direct way that heterogeneous constructionist interpretation might influence science productively. Instead of relying on some second party to do the reconstruction, could scientists—or indeed any researchers—interpret their own heterogeneous webs? Could researchers reflect explicitly on how their own social embeddedness or situatedness affects their ability to study the situations that interest them? Could they attempt to identify multiple potential sites of engagement and change for themselves? If so, this would cut through some of complexities arising from interpreters trying to model practical reflexivity.

Mapping, Map-makers, and Maps

To explore this possibility with a number of ecologists and natural resource researchers, I convened two “mapping workshops”—the first in Helsinki, co-led with ecologist-philosopher Yrjö Haila; the second in Berkeley. These workshops were designed to proceed as follows. Each researcher would focus on a key issue—a question, dispute, or action in which the researcher was strongly motivated to know more or act more effectively. All researchers would identify “connections”—things that motivated, facilitated, or constrained their inquiry and action. These might include theoretical themes, empirical regularities, methodological tactics, organisms, events, localities, agents, institutional facilities, disputes, debates, and so on. Researchers would then draw their “maps”—pictorial depictions employing conventions of size, spatial arrangement, and perhaps color that allow many connections to be viewed simultaneously. The map metaphor was meant to connote not a scaled-down representation of reality but a device that shows the way—a guide for further inquiry or action (Taylor and Haila 1989; Taylor 1990).
Over a series of sessions the workshop participants would present these maps and be questioned by other participants. As a result they might clarify and filter the connections and reorganize their maps so as to indicate which connections were actually significant resources. The ideal was that researchers would self-consciously modify their social situations and their research together, perhaps in collaborations formed among the workshop participants. Of course, given that mapping was an experiment, it was not surprising that the ideal was not realized in these initial two workshops.
Three maps from the workshops illustrate the map making that resulted. Figure 1, by a Finnish ecologist I will call “E,” was the most orderly of the maps, having been streamlined and redrawn on a computer. As such it does not do justice to the real-time experience of its production during an actual workshop. Indeed, when viewed on their own all the maps appear schematic; valuable history, emphasis, and substance were added when the mapmakers presented their maps to other workshop participants.

Figure 1 Redrawn outline of E’s map about how to conduct research on the ecology of carabid beetles in the city of Helsinki (from Taylor and Haila 1989)

The central issue on E’s map is very broad, namely, to understand the ecology of carabid beetles living in the leaf litter under trees in urban environments. On the map below this issue are many theoretical and methodological sub-problems, which reflect the conventional emphasis in science on refining one’s issue into specialized questions amenable to investigation. Above the central issue are various background considerations, larger and less specific issues, situations, and assumptions that either motivated work on the central issue or were related to securing support for the research. E’s research alone would not transform the urban public into recognizing that “nature is everywhere—including in the cities,” but by combining the upward and downward connections, he reminded himself that work on the background issues, not only refining a working hypothesis, would be necessary to be able to keep doing his research.
In narrating his map, E mentioned some additional history. Many of the ecologists with whom he collaborated had been studying a forest area, but the group lost their funding when the Forestry Department asserted that forest ecology was their own domain. It did not matter that animals are barely mentioned in the ecology of forestry scientists. The ecologists self-consciously, but of necessity, turned their attention to the interconnected patches of forest that extend almost to the center of Helsinki, and explored novel sources of funding and publicity, including a TV documentary. The upward connections were thus a recurrent, if not persistent, influence on E as he defined his specific research questions.

Figure 2 Extract from R’s map concerning research on the peasant economics and politics and tropical forest destruction in Mexico (from Taylor 1990)

Historical background depicted in a narrative format is more evident in a large map by “R,” a Mexican who had come to specialize in the economic and agronomic dynamics which lead to impoverishment of peasants, their migration into forest areas, and subsequent clearing of those forests. Figure 2 is only one section of that map. Although radically different from E’s redrawn map, R’s map also highlighted simultaneous issues of building the disciplinary and collaborative context in which to pursue his many concerns. As a biologist he wanted to stem rainforest destruction; as a political activist he wanted to reduce rural poverty; and as a resource economics graduate student in the U.S. he needed to frame technical questions that could be answered.

Figure 3 M’s map of his research into ecological degradation and impoverishment among nomadic pastoralists in West Africa (from Taylor 1990)

In Figure 3 “M,” an American studying land degradation and impoverishment among nomadic pastoralists in West Africa, depicted a more conventional conception of research. Questions form the bulk of the map and are separated from methods—the strip along the bottom. M omitted the movements, arrangements, alliances, and negotiations he built in order to monitor milk production, elicit from the herders rules governing herd movement, assess herd ownership, measure the effect of grazing on pasture growth, complete surveys to “ground truth” satellite images, and so on. M’s map also located him in his remote field area, and omitted the audiences in the U.S.—sponsors and critics alike—for his current and future research. In short, notwithstanding the guidelines I had given to mapmakers, M included the situation he studied and left himself out.
To what extent, recalling the goals of mapping workshops, did the workshops lead participants to “clarify and filter the connections and to reorganize their maps”? It took considerable time to prepare maps, and the mapmakers did not devote more time to redraw their maps in response to interaction during the mapping sessions. M, for example, did not redraw his map to include his own context. To what extent then did researchers realize the ideal of “self-consciously modify[ing] their social situations and their research together, perhaps in collaborations formed among the workshop participants”? Several participants, at the Helsinki workshop in particular, claimed that the mapping workshop had expanded the range of influences, both theoretical and methodological, that they would bring into planning their future work. One workshop participant commented that mapping made it impossible “simply to continue along previous lines.” Nevertheless, although the workshops provided the opportunity to link up with others around revealed affinities, no new coalitions emerged; changes in the researchers’ work were not so dramatic.

Extracted from Taylor, P.J. (2005) Unruly Complexity: Ecology, Interpretation, Engagement (U. Chicago Press), chapter 5, Part B.


Taylor (1990). “Mapping ecologists’ ecologies of knowledge.” Philosophy of Science Association 2: 95-109.

Taylor and Y. Haila (1989). “Mapping Workshops for Teaching Ecology.” Bulletin of the Ecological Society of America 70(2): 123-125.

Constructionist interpretations of science

In its broadest sense, the claim that scientific knowledge is constructed amounts to saying that it is not simply drawn from nature.  According to a minimal definition of constructionism, what counts as knowledge is contingent on the scientific method or framework used.  A more inclusive, social constructionism views knowledge as contingent on the scientists establishing (or disputing) the knowledge and then, through them, on their social context (see Collins’ 1981 “stages in the empirical programme of relativism.”)  During the 1980s, social constructionism (or constructivism) became the major perspective in social studies of science.  Active debate ensued about different interpretations of construction, the degree to which it is a social process, how much that process affects what counts as knowledge in the long run, the emphasis on knowledge alone or as integrated into scientific activity more generally, and whether social studies of science now needs to move beyond constructionism.  See, e.g., the exchanges among Collins, Yearley, Latour, and Woolgar in Pickering (1992) and between Sismondo (1993a, 1993b) and Knorr-Cetina (1993); Smith’s (1997) discussion of Kitcher (1993); and other exchanges reviewed in Hacking (1999).

The sense of construction as building—the process, not the product—remained underdeveloped through these debates (as noted also in Hacking 1999, 49ff).  If the project of interpreting science is construed very broadly as addressing what it means practically for agents to modify scientific activity, then the term construction has the apt connotation of a process of agents building from a number of different components.  This is the sense I explore under the label heterogeneous construction.  A similar term, heterogeneous engineering, was introduced in Law (1987).  Indeed, since the middle 1980s the literature interpreting science has included many rich descriptions of the diversity of things scientists do and use in the process of making science: scientists employ or “mobilize” equipment, experimental protocols, citations, the support of colleagues, the reputations of laboratories, metaphors, rhetorical devices, publicity, funding, and so on (Latour 1987; Clarke and Fujimura 1992, 4-5).

Nevertheless, in descriptions—even the least colored ones—there are causes or “becauses” implied in the selection and juxtaposition of different factors.  At this level of explanation heterogeneous constructionism has not been well developed.  Taylor (2005, Chapter 4) explores heterogeneous constructionism as an explicitly explanatory project, one that addresses the challenge of analyzing which of the diverse things mobilized make a difference and how they are combined to do so.

Constructionists of various stripes (Sismondo 1993a; Hacking 1999) have offered explanations of scientific developments.  Most explanations have, however, tended to move away from heterogeneity and from the implications that construction is a process (Taylor 1995).  Both these tendencies are especially evident when discussions associate social constructionism with society, context, ideology, and so on—something external to science—determining, penetrating, or being reflected in the content of accepted scientific theories.  (The diagrams in Wise 1988 provide an explicit instance of this.)  The resulting science then corresponds to the society in which it is generated or accepted.  Conversely, when “realists” or anti-relativists dispute that the outcome of scientific activity—established knowledge—corresponds to society, they hold that it must then correspond—at least, generally or eventually—with nature or reality.  The agency of scientists is not significant in these correspondence relationships.  In the former, scientists can be seen as ciphers for society or dupes for interests; in the latter, they can be forgotten once they have helped establish the knowledge.

Admittedly, published work is usually subtler than seminar or bar room discussions of social construction.  The literature generally presents the society-science relationship as refracted and allows for the observation that not all of social group X believe Y and not all believers of Y come from social group X.  In this vein we can see, for example, that scientists produce and judge knowledge mostly according to how it furthers goals of their social group (Shapin 1982).  In short, agents are quite active and practically engaged.  Nevertheless, construction in the simpler, correspondence sense has not been banished in such accounts.  If we ask how these accounts explain why this knowledge was accepted and not that, and how this knowledge was generated in the first place, the implicit “because” seems more often than not one of correspondence between knowledge and interests (Woolgar 1981; Pickering 1993).

The conceptual resources needed for heterogeneous constructionist explanations should be sought in thinking about process and practice, and not drawn from discussions of a reflection or correspondence relationship, however refracted or approximate that relationship is (Taylor 2005, chapter 4).  Moreover, so as not to perpetuate the privileged status established knowledge has had in science studies, the separation of knowledge from scientific work and activity needs to be dissolved (Clarke 1991).  Finally, the actions of agents who heterogeneously construct can never be governed solely or predominantly by the pursuit of any unitary goal, whether that goal is revealing the nature of some underlying reality (truth), establishing instrumentally reliable knowledge (Boyd 1991, 207; Hacking 1983), furthering the interests of the agent’s social group, or maximizing and concentrating the agent’s social resources (Latour 1988a, 160).  The challenge—one shared with social theory—is to develop accounts of agency in terms of widely distributed causality (see Taylor 2005; Chapters 4-6).

Extracted from Taylor, P.J. (2005) Unruly Complexity: Ecology, Interpretation, Engagement (U. Chicago Press).


Boyd, R. (1991). “On the current status of scientific realism,” in R. Boyd, P. Gasper and J. D. Trout (Eds.), The Philosophy of Science. Cambridge, MA: MIT Press, 195-222.

Clarke, A. (1991). “Social worlds/arenas theory as organizational theory,” in D. R. Maines (Ed.), Social Organization and Social Process: Essays in Honor of Anselm Strauss. New York: Aldine de Gruyter, 119-158.

Clarke and J. Fujimura (1992). “What tools? Which jobs? Why right?,” in A. Clarke and J. Fujimura (Eds.), The Right Tools for the Job: At Work in Twentieth-century Life Sciences. Princeton: Princeton University Press, 3-44.

Collins (1981). “Stages in the empirical programme of relativism.” Social Studies of Science 11: 3-10.

Hacking, I. (1983). Representing and Intervening. Cambridge: Cambridge University Press.

Hacking (1999). The Social Construction of What? Cambridge, MA: Harvard University Press.

Kitcher, P. (1993). The Advancement of Science: Science without Legend, Objectivity without Illusions. New York: Oxford University Press.

Knorr-Cetina, K. (1993). “Strong constructivism – from a sociologist’s point of view: A personal addendum to Sismondo’s paper.” Social studies of science 23(3): 555-563.

Latour, B. (1987). Science in Action: How to Follow Scientists and Engineers through Society. Milton Keynes: Open University Press.

Latour (1988). “The politics of explanation: an alternative,” in S. Woolgar (Ed.), Knowledge and Reflexivity: New Frontiers in the Sociology of Knowledge. London: Sage, 155-176.

Law (1987). “Technology and heterogeneous engineering: The case of Portugese expansion,” in W. E. Bijker, T. P. Hughes and T. J. Pinch (Eds.), The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology. Cambridge, MA: MIT Press, 111-134.

Pickering, A. (Ed.) (1992). Science as Practice and Culture. Chicago: University of Chicago Press.

Shapin, S. (1982). “History of science and its sociological reconstructions.” History of Science xx: 157-211.

Sismondo, S. (1993a). “Some social constructions.” Social Studies of Science 23(3): 515-553.

—— (1993b). “Response to Knorr Cetina.” Social Studies of Science 23(3): 563-569.

Smith, B. H. (1997). “Microdynamics of incommensurability: Philosophy of science meets science studies,” in B. H. Smith and A. Plotinsky (Eds.), Mathematics, Science, and Post-classical Theory. Durham, NC: Duke University Press, 243-266.

Taylor (1995). “Co-construction and process: a response to Sismondo’s classification of constructivisms.” Social Studies of Science 25: 348-359.

Wise, N. (1988). “Mediating machines.” Science in Context 2(1): 77-113.

Woolgar, S. (1981). “Interests and explanation in the social study of science.” Social Studies of Science 11: 365-394.

Heterogeneous construction, a variant of intersecting processes

Heterogeneous construction is a variant of the idea of intersecting processes, which, in turn, is an attempt to discipline without suppressing the unruliness of complexity.  The concept of heterogeneous construction can be illustrated by reference to the previous post on the Brown and Harris/Bowlby account of class, family, and psychology combining to explain the onset of serious depression.

As stated in the previous post, the factors are not separate contributing causes, like spokes on a wheel, but take their place in the multistranded life course of the individual [i.e., in the intersecting processes].  Each line should be interpreted as one contributing causal link in the construction of the behavior. The lines are dashed, however, to moderate any determinism implied in presenting a smoothed out or averaged schema; the links, while common, do not apply to all women at all times, and are contingent on background conditions not shown in the diagram.

For example, in a society in which women are expected to be the primary caregivers for children (a background condition), the loss of a mother increases the chances of, or is linked to, the child’s lacking consistent, reliable support for at least some period. Given the dominance of men over women and the social ideal of a heterosexual nuclear family, an adolescent girl in a disrupted family or custodial institution would be likely to see a marriage or partnership with a man as a positive alternative, even though early marriages tend to break up more easily. In a society of restricted class mobility, working-class origins tend to lead to working-class adulthood, in which living conditions are more difficult, especially if a woman has children to look after and provide for on her own. In many such ways these family, class, and psychological strands of the woman’s life build on each other. Let us also note that, as an unavoidable side effect, the pathways to an individual’s depression intersect with and influence other phenomena, such as the state’s changing role in providing welfare and custodial institutions, and these other phenomena continue even after the end point, namely, depression, has been arrived at.

Suppose now, quite hypothetically, that certain genes, expressed in the body’s chemistry, increase a child’s susceptibility to anxiousness in attachment compared to other children, even those within the same family.  Suppose also that this inborn biochemistry, or the subsequent biochemical changes corresponding to the anxiety, rendered the child more susceptible to the biochemical shifts that are associated with depression.  It is conceivable that early genetic or biochemical diagnosis followed by lifelong treatment with prophylactic antidepressants could reduce the chances of onset of severe depression. This might be true without any other action to ameliorate the effects of loss of mother, working-class living conditions, and so on.  There are, however, many other readily conceivable engagements to reduce the chances of onset of depression, for example, counseling adolescent girls with low self-esteem, quickly acting to ensure a reliable caregiver when a mother dies or is hospitalized, making custodial institutions or foster care arrangements more humane, increasing the availability of contraceptives for adolescents, increasing state support for single mothers, and so on.  If the goal is reduction in depression for working-class women, the unchangeability of the hypothetical inherited genes says nothing about the most effective, economical, or otherwise socially desirable engagement—or combinations of engagements—to pursue.  Notice also that many of these engagements have their downstream effect on depression via pathways that cross between the different strands.  For example, if self-esteem counseling were somewhat effective then fewer unwanted pregnancies and unsupportive partnerships might be initiated; both effects could, in turn, reduce the incidence of single parenthood and difficult living conditions.

These sequences of multiple causes, building on each other over the individual’s life history, permit a number of conclusions about the nature-nurture debate:

1.  Neither the unchangeability of genes nor the reliability of some gene- or biochemistry-based intervention, such as the hypothetical prophylactic antidepressants, would prove that the genes are the most significant cause of the acute depression that has been occurring in the absence of such treatment.

2.  Critics of genetic explanations could dismiss the attribution of an individual’s behavior to genes (or 50% or 80% to genes) as a technically meaningless partitioning of causes without placing themselves at the other pole from genetic determination.[i] That is, they would not have to make the counterclaim that the environment determines behavior or that, if the right environment were found, any desired behavior could be elicited. The Brown-Harris-Bowlby (BHB) account addresses malleability or immalleability of behavioral outcomes without ruling out genetic contributions.

3.  Similarly, critics would not need to rest their case on demonstrations that behavioral genetics has been or still is methodologically flawed (Lewontin et al. 1984), on textual deconstructions of the categories and rhetoric employed (Lewontin 1979), or on attributions of political bias to the supporters of behavioral geneticists. These are all interesting, but, in light of the BHB account of the behavior, not necessary for a conceptual critique of genetic determinism.

Over and above these conclusions, the BHB account of the origins of acute depression in working-class women also displays the following features that I associate with the idea that something is “heterogeneously constructed,” or an outcome of “intersecting processes.”

a) Without any superintending constructor or outcome-directed agent,

b) many heterogeneous components are linked together, which implies that

c) the outcome has multiple contributing causes, and thus

d) there are multiple points of intervention or engagement that could modify the course of development. In short,

e) causality and agency are distributed, not localized. Moreover,

f) construction is a process, that is, the components are linked over time,

g) building on what has already been constructed, so that

h) it is not the components, but the components in linkage that constitute the causes. Points c) and f–h) together ensure that

i) it is difficult to partition relative importance or responsibility for an outcome among the different types of cause (e.g., 80% genetic vs. 20% environmental). Generally,

j) there are alternative routes to the same end, and

k) construction is “polypotent” (Sclove 1995), that is, things involved in one construction process are implicated in many others. Engaging in a construction process, even in very focused interventions, will have side effects. Finally, points f) and k) mean that

l) construction never stops; completed outcomes are less end points than snapshots taken of ongoing, intersecting processes.

I am aware that there may be objections to the case I have chosen to make the preceeding points.  In discussing depression among working class women, rather than in other groups, I could be seen as perpetuating a male, professional class perspective.  However, the politics of the case can be viewed quite differently.  Although depressed working class women are the focus, the intersecting processes account brings a range of other agents into the picture.  While the account does not identify ways to cure the women studied, other girls and women that follow them might seek support from, or find themselves supported by—to pick up on the potential engagements mentioned earlier—counsellors, hospital social workers, people reforming custodial institutions, family planning workers, social policy makers, and so on.  Moreover, these agents can view their engagement as linked with others, not as a solution on its own.  For example, when women’s movement activists create women’s refuges as a step away from living in unsupportive households, this makes it possible for therapists who specialize in the psychological dynamics of the woman in her family to consider referring women to refuges as a critical disruption to the family’s dynamic.  The politics of highlighting different kinds of causes and their interlinkages can be seen as promoting such exchange among the distributed set of agents and contributing to the potential re-formation of the social worlds intersecting around the development of any given focal individual or outcome.

Extracted from Taylor, P.J., “Distributed agency within intersecting ecological, social, and scientific processes,” pp. 313-332 in S. Oyama, P. Griffiths and R. Gray (Eds.), Cycles of Contingency: Developmental Systems and Evolution.  Cambridge, MA: MIT Press, 2001.

[i] The nonpartitionability of different kinds of biological and social causes, given the interdependence of their effects, is demonstrated well by Lewontin (1974), when he argues that statistical partitioning of effects (“analysis of variance”) does not constitute an analysis of causes. Of course, partitioning of biological and social causes does have ideological meaning (Lewontin et al. 1984).


Lewontin, R. C. (1979). Sociobiology as an adaptationist program. Behavioral Science, 24, 5-14.

Lewontin, R. C., Rose, S., & Kamin, L. J. (1984). Not in our genes: Biology, ideology and human nature. New York: Pantheon.

Sclove, R. (1995). Democracy and technology. New York: Guilford.