The challenge of interdisciplinarity: metaphors, reductionism, and the practice of interdisciplinary research (child development)



Go to Google and type in ‘interdisciplinary’ as a search word. What do you get? In the first instance, the answer is almost 1.8 million entries or ‘hits.’ Not quite as many as for  W. Bush at almost more than 3.4 million hits or Manchester United at just 2 million, but nevertheless an impressive number. Combining ‘interdisciplinary with ‘psychology’ delivers over 360,000 entries, 20.2 percent of the total number for ‘interdisciplinary alone, and noticeably more (in descending order) than for ‘sociology,’ ‘anthropology,’ ‘developmental biology,’ and ‘behavior genetics.’ Within psychology, ‘social psychology’ results in many more hits than, for example, ‘cognitive psychology’ and ‘developmental psychology’ when combinations with ‘interdisciplinary’ are made. Nevertheless, each one provides an imposing numerical outcome. Repeating the whole exercise with ‘interdisciplinary research’ and ‘interdisciplinarity’ does little to alter by very much any of these relative comparisons (Table 1).

At first flush, this trawl through the Internet would seem to suggest that interdisciplinarity is well established in some areas of study represented in this volume. Unfortunately, the quantitative findings do not tally with qualitative considerations. Why not? First of all, because there is a lack of clarity about the meaning of interdisciplinarity or what constitutes interdisciplinary research. Further confusion is engendered when attempting to distinguish among interdisciplinarity, cross-disciplinarity, multidisciplinarity, and transdisciplinarity. Yet we now appear to be in the age of the inter-discipline prefixes and suffixes, with proliferations of bio-, etho-, psycho-, and socio-, together with the recent arrival of scientific endeavors dubbed ‘social neuroscience’ and ‘neuroeconomics.’ As for ‘child development,’ the number of Google entries is relatively large (Table 1). Once again, however, the numbers game masks a range of different designations as to the meanings of interdisciplinarity and interdisciplinary research. Certainly, interdisciplinarity has had something of a bad press in the past.

The up and downs of interdisciplinarity

If it appears that something of an interdisciplinary Zeitgeist is upon us, it has been achieved in the face of some strong pockets of resistance in the past. One example is epitomized by the remark of Leslie A. Smith (1900-1975) in his topic The Science of Culture (1949) to the effect that cultural anthropologists “… have sold their culturological birthright for a mess of psychiatric pottage” (p. xix). During the 1960s, some leading biologists opposed what they saw as the threat of their discipline being reduced to the laws and principles of physics, or more specifically to classical mechanics. The same mistrust is still evident in attempting to preserve disciplinary boundaries (e.g., that between psychology and neuroscience).

Why then has interdisciplinarity (ID) become the mantra of current scientific policy? Before getting anywhere near answering that question, we need to address a number of converging issues: the meaning of ID relative to cross- and multidisciplinarity as well as to transdisciplinarity, levels of (biological) organization and the associated problem of reductionism, and the use of metaphors and other tropes (e.g., analogy) in science more generally. What follows is essentially a personal view derived from the experience of being a member of so-called interdisciplinary programs of research in child development. Undoubtedly, this view will have its dissenters, particularly with regard to the restricted meaning accorded to ID. Such an imposition should be seen as a debating point, rather than a firmly held belief as to how interdisciplinary research (IDR) should be construed. The hope is that it will highlight some of the structures and processes needed for IDR in child development that go beyond mere cross-disciplinarity and multidisciplinarity.

The discipline of interdisciplinarity

In 1996, the final report of the US Gulbenkian Commission on the Reconstruction of the Social Sciences was published. While favorably disposed to IDR, it did little more than recommend it could be achieved by granting academics tenure in two departments. Nowhere in the report was there a systematic attempt to distinguish ID from the other three similar terms. In short, among other things, it is a shared language (or what might be termed a scientific Esperanto) between the participating disciplines that embraces both theory and method (Table 2).

With the establishment of such a linguistic ‘trading zone’ at the frontiers of disciplines, the task of dissipating barriers to ID has begun. If this first step is seen as a ‘mission impossible,’ there are examples in science to suggest otherwise. For instance, the interdiscipline of biophysics was established through the combined efforts of physicists, biochemists, and computer scientists to learn each other’s theoretical vocabulary in order to gain fresh insights into biomolecular mechanisms involving, for example, protein synthesis in membranes. Nearer to home, cognitive neuroscience arose from a lack of models in clinical neuropsychology that could be used to address the effects of focal brain injuries. During the 1960s, such models were sought in cognitive psychology, with the result that the neuropsychologists began to share the language and methods of cognitive psychologists.

Even more germane were the efforts of Arnold Gesell and Myrtle McGraw in the 1930s and 1940s to found the study of child development on principles drawn from embryology and particular branches of physics such as thermodynamics. Other pertinent examples are: the birth of biochemistry through Francois Magendie (1783-1855) bringing together organic chemists and physiologists to study collectively the relevance of nitrogen for animal nutrition, and the way in which Walter Nernst (1864-1941) and collaborators integrated what was then known about electrochemistry with thermodynamics during the early 20th century to give birth to what is now inorganic chemistry.

To label a scientific activity as an ostensive example of IDR is a common occurrence and a source of some obfuscation. IDR can take on at least three types, with, for example, one discipline coming to subordinate the others brought together to address a common problem beyond the bounds of a single discipline. Once more, what makes a distinction is a commonly shared language that ‘cracks’ the linguistic codes of the participating disciplines (Table 3).

Table 1. Approximate number of Google entries for interdisciplinary, interdisciplinary research, and interdisciplinarity. These terms are then combined with psychology, followed by doing the same for developmental, cognitive, and social psychology. The procedure is repeated for what might be regarded as ‘sister’ disciplines (sociology; anthropology), for two others that have a bearing on theorizing and research in developmental psychology (developmental biology; behavior genetics), and for child development.

Search word


Interdisciplinary research


On its own








Developmental psychology




Cognitive psychology




Social psychology












Developmental biology




Behavior genetics




Child development




If only it were that simple. For example, disciplines can share identical words, but they can have contrasting meanings in each one. Examples include different interpretations of growth and individuation across the developmental sciences and even that pertaining to causality. when one gets down to this level of discussion, proposed IDR projects can eventuate in disarray and the loss of a common cause. The interdisciplinary gap widens instead of closing.

Bridging the gap: levels of organization and reductionism Levelism

One way in which disciplinarity is portrayed is to arrange disciplines along a hierarchy of levels of organization and then at each level to pigeon-hole them under ‘structure,’ ‘function,’ and ‘evolution.’ Table 4 depicts such a hierarchy for the life sciences, broadly defined.

It should be evident that the number of levels and how they are labelled is, together with the disciplines included, an arbitrary exercise (e.g., ecology could have been allocated to the top and particle physics to the bottom of the hierarchy). Nevertheless, one person’s hierarchy looks very much like another’s demarcation of levels and assignment of disciplines. What is this stratified hierarchy meant to convey? There are two responses. One is that as you move up the hierarchy, disciplines have to address increasingly complex phenomena, together with the emergence of properties not manifested at the lower levels. The other is that as you move down it, increasing explanatory power can be gained, which has led to the claim that science should be unified from the bottom up rather than top down. Whichever way you move, you are confronted with a task of almost Sisyphean dimensions, namely, climbing the slippery slopes of reductionism.

Table 2. Starting from a consideration of what constitutes disciplinary, interdisciplinarity (ID) is compared to three other forms of scientific collaboration. There is still confusion and a general lack of agreement about the meaning of ID and how it should be practiced. The defining features of ID are deliberately presented in conservative terms so as to draw distinctions with the other forms of scientific collaboration that are often taken as being synonyms. Transdisciplinarity is the most vague term used to denote cooperation between disciplines. It appears to be an attempt to get science galvanized into focusing on the provision of solutions to a variety of social and economic concerns that may be national or, more commonly, worldwide in scope (e.g., environmental pollution, and its effects on child development).


Defining features


D i sc ip l i nari ty

During the early part of the 20th century, there was a ‘drive for disciplinarity’: establishment of ‘bounded’ disciplines, with their own theories, methods, and standards of scientific rigor. Gave rise to modern-day discipline structures having their own scientific societies and accreditation committees

Until the late 19th century, disciplines as they existed were more loosely ‘bounded’ in that science was pursued as an enterprise based on a broad-ranging critical reflectivity across many areas of knowledge. Such was the case, for example, in descriptive embryology. With the ‘push for specialization,’ new disciplines were founded (e.g., pediatrics, which became a ‘bounded’ discipline in the 1930s). Largely as a result of the Cold War, area studies and systems approaches to science began to emerge in the late 1950s which ultimately gave rise to what have been termed ‘interdisciplines’ (e.g., cybernetics)

Interdisciplinarity (ID)

Well-established disciplines working together on a common problem, but with the express aim of adjusting their theories and methods so that they can be integrated into a new discipline or interdiscipline. It involves generalizing from multidisciplinary settings so that a common language covering theory and method can be established

In the past, there have been a number of unsuccessful attempts to establish a common scientific language (e.g., behaviorism; logical positivism; General system theory) and the quest continues (e.g., on a more restricted scale with the theory of embodiment). Apart from that, most individuals participating in this ‘strong’ form of scientific collaboration do so not only to contribute to another field, but also to take back new ideas to their own disciplines (thus preserving discipline independence)


Disciplines working together on a common problem, but not changing their approaches or adjusting to the knowledge base or techniques of other disciplines. Participating disciplines then tend to present their findings in discipline-dedicated conferences and journals

Most so-called ID research takes on this ‘weak’ form of scientific collaboration


Takes on two forms:

1. researchers in one discipline (e.g., physics) choose to work in another discipline (e.g., biology)1

2. researchers trained in two disciplines (e.g., psychology and neuroscience or psychology and anthropology)

Two noticeable and increasing features of modern-day science are:

1. cross-appointments between departments (e.g., between computer science and psychology)

2. cross-disciplinary training programs (e.g., within the context of the neurosciences)


A sort of half-way house between disciplinarity and ID in which the aim is to provide a forum or platform for the generation of new ideas that can then be applied across a number of disciplines

If properly understood, it seems to be a medium created so that non-scientists, can have a say in the decision-making process as to which scientific problems need to be addressed. Consequently, it tends to lead to calls for science to tackle issues such as diseases and discrimination, and to providing a better standard of living for all

1 Outstanding examples of this type of cross-disciplinarity are Max Delbruck (1906-1981) and Leo Szilard (1898-1964), both trained in quantum mechanics, who applied their knowledge acquired in physics to the study of cell reproduction. Their work made a significant contribution to the discovery of the DNA double helix attributed to John D. Watson and Francis H. C. Crick.

Table 3. Three types of interdisciplinary research, which ultimately depend on whether or not the participating disciplines share a common language, and for which possible examples involving psychology and possible common problems are given.  



Possible example

Possible common problem

Communality in vocabulary

Two or more disciplines focusing on a common problem, with a common scientific language and set of concepts and techniques as well as shared standards of rigor and proof. While a common shared language may be assumed, it could turn out that some terms have different meanings between the participating disciplines

Psychology and Behavioral biology

Development of attachment

Disparity in vocabulary

Two or more disciplines with different languages and concepts as well as techniques, and standards of proof The problem to be tackled is divided up so that each part can be dealt with by relevant disciplines. Findings from the parts then have to be integrated in some way

Psychology and Anthropology

Cross-cultural comparisons of parent-child communication

Disparate in vocabulary and subordination of one discipline to another

Two or more disciplines with very different languages, research methods techniques, and standards of proof There is a search for a common language, which requires major adjustments in concepts, methods, and techniques. The outcome can be a hierarchically arranged research strategy in which one discipline is subordinated to another in tackling a common problem.

Psychology and Pediatrics

Development of very preterm infants

Table 4. Levels of organization in relation to structure (being), function (acting), and evolution (becoming) and the (sub-)disciplines that address each one. Evolution is meant to denote the study of change over different time scales (viz., real, developmental, and geological time).






Cultural anthropology




Management science

Political science

Cultural anthropology


Social psychology

Social psychology

Developmental psychology1




Developmental psychology







Bi ochemistry



Molecular biology

Molecular biophysics

Developmental genetics

1 Developmental psychologists carry out research at this level when, for example, it involves the analysis of family dynamics

2 Neurophysiology can be interpreted as covering neuroscience and developmental neuroscience and thus can feature, for example, at both the organic and cellular levels under ‘Evolution’


Here is not the place to embark on a detailed diatribe about the provenance of reductionism in science in general and for IDR in particular, and which assumes not one, but a number of slippery slopes. Instead, we focus just on theoretical reductionism. To begin with, what is meant by theoretical reductionism?

Termed intertheoretic reductionism by Churchland (1986), it concerns the explanation of the reduced theory (e.g., the theory of gases) by the reducing theory (e.g., statistical mechanics). On a grander scale, it encompasses the pursuit of a Theory of Everything as strived for by General system theory in the past and by such as string theory, superstring theory, and M-theory at present. In the context of the deductive-nomological model of scientific explanation originating with Carl Gustav Hempel (1905-1997) and Paul Oppenheim in 1948, theoretical reductionism is supposed to work through the implementation of bridge laws or principles. These devices act as transformation rules for linking two distinct linguistic expressions with two theories at different levels. Self-organization is sometimes treated as possessing the potential to become a bridge law as are Piaget’s functional universals (viz., assimilation, accommodation, and equilibration). The problem with bridge laws is that they can become too cumbersome to put into practice such that they defeat the purpose of ever attempting theoretical reductionism in the first place (a case in point being the way in which Piaget attempted to operationalize equilibration). If this is so, and which appears to be borne out by the fact that the most successful reductions in the history of science (e.g., of Mendelian to molecular genetics) did not have recourse to bridge laws, then an alternative strategy is needed.

If not bridge laws, then what? Let’s put this question to one side for a minute and consider two classic problems of theoretical reductionism. These are genetic determinism and the relationship between psychology and neuroscience.

1. Genetic determinism: with the success of the Human Genome Project, there is an increasing tendency to regard genes as the ultimate determinants of development and of developmental disorders. Knowing the sequence of many human genes, however, is not going to be particularly revealing about development, given the protein-folding problem and continuing ignorance of the pathways between genotype and phenotypes during development. Genetic determinism brings with it the danger of reification: reducing something that is a dynamical process to a static trait and then searching for its single (genetic) determinant. Examples include aggression, intelligence, and syndromes such as ADHD. Without doubt, genes influence virtually all behavior, but virtually no behavior is determined by them. Structural genes manufacture proteins and enzymes whose translation and regulation are critical to phenotypical changes in ontogenetic development (and biological evolution). However, the environment can inject some degree of developmental specificity as well (e.g., the sex of a turtle depends on the temperature of incubation and not on the dictates of chromosomes). In this example, the environment is instructive and the genotype permissive.

2. Psychology and neuroscience: without doubt, one of the most enduring themes in the history of science is how to conflate psychology and neuroscience into a unified theory of behavior or cognition. Can psychology be reduced to neuroscience as some contend (Churchland, 1986)? Or is neuroscience irrelevant to psychology as maintained by others who see their task as defending the autonomy of psychology from intrusions by other sciences (Fodor, 1975)? The nub of the issue is whether mental states (e.g., emotions, feeling, and consciousness more generally) can be reduced to corresponding neural states. Recent attempts that have been made to resolve this issue include Gerald Edelman’s theory of neuronal group selection. Churchland’s (1986) response, in a pro-reductionist mode, has been to argue that a psycho-neuro symphysis can be achieved by what she calls theoretical co-evolution: theories at different levels may co-evolve such that they inform and correct each other, thus bringing them ever closer to assuming a common theory. As Churchland herself realizes, while concordant development has worked for the marriage of thermodynamics and statistical mechanics as well as for physics and chemistry more generally, there are still formidable problems to be overcome in fusing psychology with neuroscience. Why? Because it is still unclear how knowledge of the brain exerts constraints on theorizing about psychological functions. Ultimately, clarity can only be achieved through further insights into structure-function relationships. For developmental psychology, understanding such constraints seems at best remote given the ever-changing relationships between structure and function during development. Thus, psycho-neuro IDR concerned with child development faces considerable hurdles, not just because of linguistic disparities between the two fields of study (Table 3), but rather due to the lack of a common theory that goes beyond correlating changes in structure and function.

So, if not bridge laws, then what? An alternative to such laws is the use of analogies to connect two or more different levels of organization. Perhaps the most frequently cited example of the value of analogies in promoting scientific advancement is how Darwin arrived at his theory of natural selection. To begin with, he drew an analogy between artificial selection as used by animal and plant breeders and the process of natural selection. He then addressed another analogy, namely, that between the theory of population pressure developed by Thomas R. Malthus (1766-1834) and the process of speciation. In combining these two analogies, Darwin created the very foundation of modern biology.

If analogical reasoning worked as a first step for Darwin, then we can ask if it serves the same function in getting IDR off the ground (i.e., whether it provides a starting point for the development of a common language). Asking this question raises the more general issue of the role of tropes in science. To begin with, let’s take a trip to Milton Keynes.

Headline news: “Milton Keynes is to double in size over the next 20 years” (Guardian newspaper, January 6,2004)

Metaphor, analogy, and homology

Milton Keynes (MK), like Basildon, is one of the so-called new towns built in the UK during the late 1940s. Apart from having the longest shopping mall in the world according to the Guinness topic of Records, it was built on a grid network system of roads and is now home to a range of light industries. Doubling its size will make it comparable to Pittsburgh in terms of the number of inhabitants. One of these inhabitants might say:

1. MK is paradise on earth

2. Although designed differently, MK has the same functions as Basildon, which also has a number of light industries

3. Although both have a grid system, MK has different functions than Pittsburgh, with its traditional base of heavy industries

Admittedly, these comparisons stretch credulity a bit, but they do raise some relevant points. What are these points? They are that:

1. is a metaphor (note it is not a simile as our inhabitant would have said: “MK is like paradise on earth”)

2. is an analogy (viz., two different structures have similar functions)

3. is a homology, which is not a trope (viz., two corresponding structures have different functions). Relatedly:

4. Asking whether MK will have the same structures or functions in 2024 as now is a question about serial homology (viz., with development or evolution, whether or not organisms retain the same structures or functions).

A metaphor is a figure of speech in which an expression about an object or action is used to refer to something it does not literally denote in order to suggest a similarity. It is one of two master tropes, with analogy being a sub-class of metaphors. To complete the picture, the second master trope is metonymy, with synecdoche as a sub-class.

Like a metaphor, an analogy is a linguistic device or form of reasoning that logically assumes that if two things agree in some respects (mainly their relations), then they probably agree in others. To this extent, an analogy is regarded as an extended metaphor or simile. And like a ‘metaphor,’ it gives insights into the unfamiliar and unknown by comparison with something familiar and known. Furthermore, analogies are made explicit by similes and are implicit in metaphors. In practice, it is hardly feasible to delimit the use of metaphors, analogies, and similes in science. Thus, for the time being, these tropes will not be distinguished further, with the term ‘metaphor’ being used for all three.

Aristotle (384-322 BP) in his Poetics stated that the greatest thing by far was to be master of the metaphor and that to have achieved mastery is a sign of genius. A bit of an overstatement perhaps, but it is widely accepted that the functions of metaphor are indispensable to science, with a minority who think otherwise. Its acknowledged functions are: aids to communication, resources for the discovery of novel insights and the generation of new theories, and in applying a theory to data by means of metaphorical redescription (i.e., in mediating its application to real-life phenomena). Examples abound, across many branches of science, about the theory-invigorating properties of metaphors (Table 5).

Having championed metaphors as a first-staging post in implementing IDR, it is well to consider what has been said about their limitations. In short, according to some, there is a price to pay for using metaphorical identifications (Table 6). Despite such pitfalls, it is questionable whether there can be a metaphor-free knowledge of whatever phenomenon we are striving to explain.

What about homologies? What role, if any, can they be accorded in IDR? Posing this question brings in its wake the more general concept of isomorphisms between levels of organization.

Homologies and isomorphisms

While homology is one of the most important concepts in biology, it is used for quite different purposes (e.g., some morphologists define homology with reference to a common developmental origin and although a different concept it is sometimes the case that the two homologies can be congruent). In evolutionary biology, it stands for correspondences between species in parts of morphological structure, a segment of DNA, or an individual gene. It becomes controversial when applied to behavior and development. Why? Because, in principle, homology is a qualitative concept (viz., something is homologous or not) and thus it can only be applied with considerable difficulty to phenomena that show a great deal of variability such as behavior and development. Despite this problem, there are ongoing attempts to convert homologies into mathematical isomorphisms and to account for development in terms of serial homologies.

Table 5. Examples of theories and concepts that emerged from particular metaphors (or analogies) in terms of who used them (‘Source’), where they came from originally, and to what field of study they were applied. Freud and Piaget are renowned for their use of metaphors in generating their respective theories. James Clerk Maxwell was openly honest about the sources of his metaphors and another one who used them widely in his work.? = Could it have been Aristotle?





Theory of natural selection


Animal breeding

Evolutionary biology

Theory of electricity and magnetism


Fluid mechanics

Electromagnetic fields

Epigenetic landscape








Assimilation and accommodation


Digestive system functioning

Genetic epistemology/Developmental psychology





Table 6. Three problems put forward as being associated with the use of metaphors (and analogies) in science. Lewontin’s metaphorical distortion is by far the most problematic.




Misplaced metaphor or Lavoisier’s problem

Proposing a metaphor that turns out to have no value in understanding the target phenomenon

Antoine L. Lavoisier (1743-1794) proposed that a living organism is like a combustion engine. While subsequently shown to be completely incorrect, it brought together chemistry and biology, thereby encouraging physiologists of the time to take account of chemistry in their work. This eventually gave rise to modern insights and formed the basis for the initial establishment of biochemistry. Thus, misplaced metaphors can lead to advances in science, even when they are shown to be wrong, by means of testing them out

Metaphorical distortion1

A theory provides explanations and a model the related analytical techniques. In applying the model to a real-world phenomenon, the latter needs to be associated with some metaphor. Such metaphorical identification can give rise to metaphorical distortion (or what others have termed ‘sort-crossing’)

An example of a metaphorical distortion is treating evolution as though it were a process of trial and error. Doing so runs the risk of imposing concepts such as ‘intention’ and ‘will’ on what is seen as generally being a random process

Overreliance on metaphors

“Major reasons for psychology’s lack of progress in accounting for brain-behavior relationships stem from a reliance on metaphorical explanations as a substitute for a real understanding of neural mechanisms”2

Such a statement is not supported by the vast literature on metaphors in general and their use in science in particular. For example, if Charles S. Sherrington (1857-1952) had not put forward his notion of a (then unseen) synapse as metaphor for neural connectivity, then S. Ramon y Cajal (1832-1934) would probably never have fully developed the neuron doctrine

1 R. C. Lewontin, 1963. Models, mathematics and metaphors. Synthese, 15, 222-244.

2 V. S. Ramachandran and J. J. Smythies, 1997. Shrinking minds and swollen heads. Nature, 386, 667-668.

The distinction between homology and analogy is embedded within the more general concept of isomorphisms. There are three sorts of isomorphisms to be drawn between different levels of organization:

1. Analogical isomorphisms: also known as the ‘soft’ systems approach, the concern is to demonstrate similarities in functioning between different levels. However, they say nothing about the causal agents or governing laws involved.

2. Homological isomorphisms: also known as the ‘hard’ systems approach, the phenomena under study may differ with regard to causal factors, but they are governed by the same laws or principles based on mathematical isomorphisms. The latter can be derived, for example, from allometry, game theory, and linear or non-linear dynamics, as well as a broad range of frequency distributions (e.g., Poisson distribution).

3. Explanatory isomorphisms: the same causal agents, laws, or principles are applicable to each phenomenon being compared.

The interdisciplinary exercise of approaching ontogenetic development as a process of interacting dynamical systems in developmental psychology has been mainly confined to (1), but it strives to attain (2), and for which there are some recent examples (e.g., in applying chaos theory to the study of how fetal and infant spontaneous movements are organized).

A serial homology addresses the issue of whether repetitive structures within the same organism are the same or different. When brought to bear on development, it results in questions such as: in what ways is behavior pattern A at time T1 the same as or different from that at T2? Are they served by homologous or analogous structures at the two ages or by those that are partially homologous and partially analogous? Such questions confront what in essence calls for IDR, namely, the evergreen topic concerning the development of structure-function relationships.

We now turn from the abstract to things more pragmatic: the practicalities of doing IDR (with the remark that the OED defines ‘pragmatic’ as dealing with things sensibly and realistically in a way that is based on practical rather than theoretical considerations).

News flash: “Pushing the frontiers of interdisciplinary research: an idea whose time has come” (Naturejobs, March 16, 2000)

This five-year-old news flash was a blurb for a number of US research initiatives that were accorded the adjective “interdisciplinary.” In particular, coverage was given to the Bio-X project housed in the Clark Center at Stanford University, which gathers together researchers from engineering, the chemical and physical sciences, medicine, and the humanities. What is the project meant to achieve? One senior academic associated with the project answered as follows: “What’s really interesting is the possibility that we have no clue what will go on in the Clark Center. That’s the point. Much of what we think works is this random collision that has a physics person talking to somebody interested in Alzheimer’s” (p. 313).

The Bio-X project, as with others of the same ilk, is an example of ‘big science,’ largely concerned with the development of new (biomedical) technologies. In its present instantiation, it is best labeled as a cross-disciplinary program of research that, perhaps with more of the random collisions, could evolve into a series of IDR projects. Certainly, it is a more expensive way of achieving truly IDR than ‘small science.’ The latter, as an ID enterprise, begins with a focus on a commonly defined problem emanating from negotiated theoretical settlement arrived at through the medium of metaphorical reasoning and the like. How small should ‘small’ be though? If forming an across-discipline group to establish guidelines for achieving desired outcomes in patient care is of any relevance, then the recommendation is not to exceed twelve to fifteen members, with a minimum of six (Shekelle et al., 1999). Too few members restrict adequate discussion and too many disrupt effective functioning of the group.

Assuming that a common problem has been identified, what are the further practical considerations to be borne in mind when attempting to carry out IDR? Some, but by no means all, can be captured under three headings: preliminary questions, having clarity about general guidelines and goals, and overcoming threats to IDR.

° Preliminary questions

1. What does IDR achieve that would not be attained by a single discipline?

2. In what ways would IDR give rise to improved and more powerful explanations?

3. What disciplines should be included and excluded (or at least held in abeyance)?

4. Does a new vocabulary interpretable by all participating disciplines need to be developed?

5. Do new methods and techniques need to be developed?

° General guidelines and goals

1. The main aim of IDR should be to predict and explain phenomena that have not been studied previously or are only partially understood and resolved.

2. Establish criteria for judging what counts as good quality IDR. As yet, there are no well-defined (i.e., operationalized) criteria for making such a judgment. On a personal note, at least one good indication that an IDR project is proceeding well is if a member of the team (e.g., psychologist) is able to report findings relevant to another from a different discipline (e.g., pediatrician) coherently at a conference mainly for colleagues of the latter.

3. The publications stemming from IDR should report not just the methods of data collection and analysis, but also how ID collaboration was achieved. Incorporating how this was done can be of benefit to others attempting to initiate IDR, as well as providing a source of reference for developing and improving the practice of such research.

4. At all costs, avoid the ‘Humpty-Dumpty’ problem: allowing participants to pursue their own discipline-related research agendas without regard to what has been defined as the common problem, such that at a later stage the pieces have to be put together to form a coherent whole. In order to prevent this:

5. Constantly ask what the common problem and related questions are in the first instance. Are we still ‘on track’ or are we losing sight of the original plan for achieving the desired outcomes? What were the desired outcomes and do we need to alter them in some way, given how things have gone?

° Threats to IDR: apart from one discipline riding atop a hierarchy of subordinated disciplines as mentioned previously, others are -

1. Continuation of research funding that endorses existing disciplinary boundaries

2. Career paths in academia continue to be dependent on discipline-best performance criteria

3. Not encouraging technical staff (the lifeblood of most research activities) to publish in their own right. However:

4. Ensuring that the research is not primarily driven by the availability of technological innovations. While the development of new techniques is a laudable goal in IDR, they can assume a life of their own in that they permit questions to be pursued across disciplines that would not otherwise be answered. The opposite of this and another threat is:

5. Technical inertia: as pointed out by Paul Galison in his topic Image and Logic (1998) for the case of particle physics, techniques, instruments, and experimental expertise can possess an inertia that determines the course of the research. And last, but not least:

6. The First Law of Scientific Motivation: “what’s in it for me?”

As a final comment on the practicalities of IDR, its defining character is to have a shared common problem that can only be addressed by two or more disciplines working closely together. In tackling it, Hodges’ Law of Large Problems has a very practical implication: inside every large problem is a small (and more manageable) problem struggling to get out.


Research in child development has long been distinguished by multidisciplinarity, if not interdisciplinarity. In the 1930s and 1940s, both Gesell and McGraw had embarked on research programs addressing core issues about the nature of infant development that were both theoretically and in practice steadfastedly committed to the ethos of interdisciplinarity. McGraw, for example, brought together an interdisciplinary team consisting of researchers from biochemistry, neurophysiology, nursing, pediatrics, physiology, and psychology as well as requisite technicians during her time at the Babies’ Hospital of Columbia University (Dalton & Bergenn, 1995b, p. 10). Her studies were sponsored by the Rockefeller Foundation, which had a special commitment to the promotion of IDR.

Times have changed and nowadays it is less common to find such an array of disciplines collectively focused on resolving a common set of problems concerning child development using a judicious interplay of cross-sectional and longitudinal methods. This is not to imply that IDR is a good thing and specialization a bad thing for research on child development. Many breakthroughs have been achieved (e.g., in studying cognitive development) from within a more or less monodisciplinary framework. IDR is mandated by the start point for any sort of research: “What’s the question?” What is at issue is whether the question, when pared down so as to render more specific ones that are methodologically tractable, unequivocally carries with it the necessity of crossing disciplinary borders.

The success of IDR depends initially on the thoroughness of attempts to develop a common language of communication framed around a common problem. Achievement of a common language should suggest isomorphisms between levels of organization representative of the disciplines involved and which emerge from the skillful use of metaphors and analogies, and perhaps ultimately homologies. The power of metaphorical reasoning to achieve communication between individuals from different backgrounds has been demonstrated, for example, in research on consultations between pulmonary physicians and their patients (Arroliga et al., 2002). If it works so successfully in this sort of setting in which such a marked disparity in language use has to be overcome, then this is surely an indication of its potential for fostering IDR.

Inevitably, reductionism in one form or another looms large in the context of IDR. Despite the rise of radical reductionism in the guise of genetic determinism during recent years, there is little evidence to suggest it has any real significance for the way in which most developmental scientists conduct their research. What one finds is that reductive analysis (i.e., induction) is combined with holistic synthesis (i.e., deduction), which have commonly (and mistakenly) been represented as mutually exclusive types of scientific explanation. Embryologists such as Paul A. Weiss (1898-1989), a staunch defender of holism, long ago argued for the necessity of maintaining both approaches in research on living systems. Put another way, it is an argument that both upward and downward causation should be accounted for in IDR.

Organizational structures need to be in place in order for IDR to flourish and in this regard the USA is still ahead of the game. On the one hand, there are agencies that continue to promote and support IDR networks, such as the MacArthur Foundation, some of which are committed to the study of child development (e.g., Network on Early Experience and Brain Development). On the other hand, there is considerable encouragement for the establishment of interdisciplinary teaching, at least with respect to the undergraduate level, through the activities of the Association for Integrative Studies. In order to overcome the confusion about the meaning of interdisciplinarity, this organization commissioned a task force whose work culminated in a report entitled “Accreditation Criteria for Interdisciplinary Studies in General Education” (2000). While a first step in identifying good practice in interdisciplinary teaching, this document also helps in removing some of the ambiguities surrounding the use of the term interdisciplinarity more generally.

Why has interdisciplinarity become the mantra of scientific policy? The optimist might answer that it is because it provides the sort of intellectual challenge that leads to scientific breakthroughs. Apart from mentioning the potential financial savings to be gained from replacing a diverse multidisciplinarity with a more unified interdisciplinarity (or in other words, amalgamating departments when there are cash-flow problems), the pessimist would point out that the policy makers have overlooked Inertial Principle: asking scientists to revise their theory is like asking a group of police officers to revise the law. Now there’s a challenge.

Theories of development

The aim of this part is to explain the main features of theoretical approaches to development that have shaped contemporary developmental sciences in general and developmental psychology in particular. The strengths and weaknesses of each approach will be indicated. The final section on the application of dynamical systems approaches to development enables further details to be added to the interdisciplinary framework outlined in the Introduction.

Next post:

Previous post: