2.4 Developmental Systems Theory

Developmental psychology has long been organized around the so-called “nature–nurture debate,” which asks how much of a behavior or characteristic is “innate” or “genetic,” and how much is due to “the environment.” This division is familiar, but it is also misleading. Development cannot be understood by separating nature from nurture and then asking how much each contributes.

Mid-twentieth-century debates about innateness were often organized around an opposition between instinct and learning. American behaviorism emphasized the ways behavior was acquired through experience, while European ethology, associated especially with Konrad Lorenz, investigated species-typical behaviors such as nest building, migration, courtship, and imprinting. Despite their differences, both traditions generally preserved the idea that behavior could be divided between what was already present in the organism and what was acquired through learning (Johnston, 2010).

Lorenz (1950) explained apparently instinctive behaviors as inherited adaptations shaped by natural selection. Experience might release or guide them, but their basic organization was understood to be already present in the organism. These cases seemed especially compelling because they appeared to show complex, organized behavior emerging without having been constructed through development.

Researchers in developmental psychobiology challenged this distinction directly. Figures such as Daniel Lehrman and T. C. Schneirla argued that “innate” should not function as an explanation of behavior, an argument that later informed Developmental Systems Theory (DST). Dividing behavior into the innate and the learned encouraged researchers to stop asking how it developed. Moving a behavior from one category to the other would not solve that problem. Even species-typical behavior has a developmental history, emerging through the activities of an organism under particular developmental conditions (Lehrman, 1953).

A central figure in this tradition was Gilbert Gottlieb (1929–2006), whose research from the 1960s onward made the problem of innateness experimentally tractable. Gottlieb worked within developmental psychobiology, an interdisciplinary approach that explains behavior through interacting biological, behavioral, social, and environmental processes (Michel, 2013). Rather than accepting apparently unlearned preferences as evidence of an inherited program, he asked what normally occurring developmental experiences had made those preferences possible. His studies of newly hatched ducklings and their attraction to the maternal call became a particularly revealing case.

An important precursor was Zing-Yang Kuo, who had opposed instinct explanations since the 1920s. Kuo showed that even the apparently instinctive rat-killing behavior of cats could be altered by changing the conditions in which kittens were reared. The point was not that cats had no species-typical organization. It was that the behavior could not be explained by naming an instinct while leaving its developmental construction unexamined (Kuo, 1930).

The experimental strategy shared by Kuo, Gottlieb, and others was to systematically alter the ordinary course of development. What appears inevitable under normal conditions may depend on resources that are themselves so reliable that they disappear from view. By changing those conditions, researchers could reveal dependencies hidden within the usual developmental pathway. An “instinctive” behavior did not simply emerge from nowhere; it was constructed through a system whose relevant activity became visible when some part of that system was changed.

Testing “Innate” Preferences

Earlier research had shown that newly hatched ducklings, before contact with a mother, moved toward a speaker playing the maternal call of their own species rather than the call of another species. Gottlieb incubated and hatched eggs in isolation and found that the preference remained (Gottlieb, 1965). Surely this showed that the preference was innate? Well, the story turned out not to be quite so simple: isolation from the mother had not meant isolation from auditory experience.

Embryonic ducklings vocalize inside the egg. During the days before hatching, the bill enters an air space and the embryo produces contact calls of its own. Gottlieb developed a procedure for devocalizing embryos before these calls began. After hatching, ducklings deprived of this normally occurring auditory activity no longer showed the usual preference for the mallard maternal call over the chicken call, while ducklings given a sham operation kept it, so the surgery itself was not responsible. Later experiments showed why: auditory deprivation broadened their responsiveness to slower repetition rates, including the rate characteristic of the chicken call (Gottlieb, 1978). A preference that had looked present in advance depended on activity generated by the developing organism itself.

Gottlieb’s (1980) study isolated the relevant feature of that experience. The mallard maternal call is delivered at about 3.7 notes per second, while the embryonic contact-contentment call occurs at about four. When devocalized embryos heard the embryonic call at this normal rate, their species-typical responsiveness to the maternal call at that rate was maintained after hatching. Exposure at other rates did not have the same effect.

This sequence of results is what makes the work more than a curiosity about ducks. Removing the embryo’s vocal activity disrupted the later preference; providing the relevant pattern of sound before hatching maintained it. The developing system had been using a resource produced by the organism itself that earlier explanations had not even included. Gottlieb’s work did not show that genes were irrelevant or that any experience would have produced the same result. It showed that a behavior appearing to express a built-in program depended on a highly specific developmental relation. If this is true of a preference that looks so thoroughly “hardwired,” the category of innateness begins to lose its explanatory force.

Developmental Systems Theory: Against Preformationism

Gottlieb’s experiments did more than complicate one claim about ducklings. They exposed the weakness of dividing behavior into what is innate and what is learned. Together with work across developmental psychobiology, they helped establish the approach that became Developmental Systems Theory (DST).

DST also takes up a question left open in DNA as a Developmental Resource. DNA is better understood as a developmental resource than as a plan or controller, but is it simply one resource among others? There are reasons for hesitation. Cells copy DNA with high fidelity and repair damage to it, preserving sequences across generations with unusual reliability. The question is not whether DNA has distinctive properties, but whether those properties justify giving it explanatory priority over the rest of the developmental system.

The fundamental stance of DST is opposition to preformationism in its modern forms. Godfrey-Smith (2000) develops this point through the idea of explanatory symmetry: genetic and non-genetic resources may differ in their properties, but neither should be treated as containing the developed outcome in advance. DST gives us no reason to assume that all developmental resources are equivalent. It asks us to justify the priority given to genes.

Any distinctive role DNA plays therefore has to be demonstrated rather than assumed. The genome is not a blueprint or program for development. Instead, it is a material resource used during the construction of the organism, alongside many other resources including the cellular environment, the uterine environment, the caregiving environment, and the wider physical and cultural context.

As in Life Cycles and the Direction of Development, “developmental system” here refers to the resources and relations, within the organism and beyond it, through which an individual life unfolds. Some DST writers use the term more broadly, so that the organism becomes one more element of the system alongside everything else. I think that goes too far. Resources are resources for something, and that something is the organism: a living process that keeps its own organization going and acts on its surroundings. A bird builds and maintains its nest, and the nest does nothing comparable for the bird. The genome has no automatic priority among developmental resources, but the organism is not one resource among the others. It is the center of the system whose development we are trying to explain.

DST also rejects the genome as an essence that defines the organism while requiring no further explanation. When genes are treated as controlling development, they become a “prime mover” whose own role no longer needs to be explained. Rejecting this view does not imply a blank slate on which anything can be written, or that an acorn planted in the desert might become a cactus. It changes the account of heredity. What recurs across generations is not a transmitted plan but a life cycle, reliably reconstructed as the relevant resources, including the genome, come together again (Griffiths & Hochman, 2015).

Some proponents of DST describe this as parity among developmental resources. DNA is unusually stable, but stability is not unique to DNA. The conditions of a nest, the composition of a uterine environment, the temperature range of a habitat, and the caregiving practices of a social group may also recur reliably across generations. They recur because the life cycle and the wider developmental system are being reconstructed. From this perspective, the special status often assigned to the genome may partly reflect that its stability is especially visible and measurable, while the stability of other resources is easier to overlook (Oyama, 2000). Parity in this sense applies among resources. It does not put the organism on the same footing as the resources through which it develops.

Developmental processes are also contingent, meaning that they depend on conditions that could have been different. The path taken so far makes some changes possible, but it does not make the next step inevitable. At the same time, development is constrained. What has already developed shapes the possibilities that remain.

The title of the collection Cycles of Contingency captures this combination of dependence and continuity (Oyama et al., 2001). What happens at one point changes the conditions for what can happen next, and those new conditions become part of the developing system. Across a life cycle, reliable forms can emerge through this history without having been specified in advance.

In this sense, development is historical not merely because earlier events occurred, but because some consequences of those events are carried forward in the organization of the system. A developmental history is therefore different from a simple sequence of events: what came before becomes part of the conditions under which what comes next can occur.

The duckling embryo did not merely receive an auditory environment supplied from outside; through its own vocal activity, it helped produce the conditions under which its later preference developed. More generally, organisms participate in their own development through movements and activities that alter what they encounter next.

Heredity and Behavior Genetics

The claim that genes should not receive automatic explanatory priority can sound modest in the abstract, yet it runs against a deeply familiar picture of the person as built upon a genetic foundation. Quantitative behavior genetics asks how variation in traits within a population can be statistically partitioned into genetic and environmental components. Its original tools were twin studies designed to estimate the heritability of behavioral traits. More recently, the field has turned toward genome-wide association studies (GWAS), from which polygenic scores are calculated to summarize statistical associations between genetic variation and physical or behavioral outcomes.

Traditional quantitative behavior-genetic models partition variation within a population into genetic and environmental components and may also estimate gene-by-environment interaction. These estimates answer a legitimate question about populations: how variation in a trait is distributed across a particular group of people living in a particular range of environments. They do not identify causes operating inside any one person, because every genetic contribution is realized through a developmental system. The trouble comes when a population statistic is read as a cause within the individual, and that is where behavior genetics and a developmental systems view part company (Turkheimer, 1998; Witherington et al., 2018).

Polygenic scores can have predictive value, but their meaning depends on the reference population, environmental context, and statistical model from which they are derived. Furrer and Turkheimer (2026) argue that a polygenic score is therefore a probabilistic population-level estimate rather than an intrinsic property residing inside the individual. Changing the reference population, context, or model can change the estimated risk associated with the same genome.

This differs from a pathogenic CAG expansion in HTT associated with Huntington’s disease, for which a well-characterized biological pathway links the variant to disease. Describing a polygenic score as a genetic predisposition therefore turns a population-level statistical construct into something that sounds like an inner essence, which is exactly the essentialist tendency at issue here.

The deeper disagreement concerns what statistical prediction can explain about development. The path from genotype to phenotype may be acknowledged as complex, but partitioning variation does not describe the developmental organization of that path. The role of any gene depends on the context in which it is expressed, and that context changes as the organism develops. From a DST perspective, the genome is one resource among many, and its contribution cannot be separated from the system in which it operates.

These positions also begin from different accounts of mechanism. One approach seeks explanation by isolating components and estimating their individual contributions. This can be informative, but it may obscure the extent to which the causal role of a component depends on the organization of the wider system. Behavior genetics often approaches genes in this way, estimating how variation in them contributes to differences between individuals. A starting point compatible with DST instead treats living organisms as organized wholes whose capacities depend on their components, while the causal capacities of those components are in turn shaped by the organization of the organism as a whole (Walsh & Wiebe, 2020).

These different accounts of mechanism affect what we expect to learn by taking a system apart. When organized interactions among components generate properties that the components do not possess in isolation, this is often described as emergence. Nothing mysterious needs to be added to the system. The point is that what a component does depends partly on its relations with other components and on its place within the organization of the whole. If development works in this way, studying components in isolation may miss properties that arise only through their organized interaction.

What Evolves?

By redefining heredity as the reliable recurrence of a life cycle rather than the transmission of a set of instructions, DST changes what needs to be explained about evolution as well as about development.

On the conventional view, what is inherited is genetic material, what evolves is a population’s genes, and development is the mechanism through which inherited genes get expressed. Development is therefore treated as something that happens between the more important parts of the story. But if what recurs across generations is a life cycle, reconstructed each time through a developmental system of resources and relations, then what evolves is the life cycle together with the system that reconstructs it, rather than a set of genes traveling through it.

If development is what reliably reconstructs the life cycle, then development is not downstream of evolution. It is where the phenotypic variation that evolution works with is generated.

Sources

Furrer, R. A., & Turkheimer, E. N. (2026). Polygenic risk scores are not genetic predispositions. Nature Human Behaviour, 10, 1177–1178. https://doi.org/10.1038/s41562-026-02475-8

Godfrey-Smith, P. (2000). Explanatory symmetries, preformation, and developmental systems theory. Philosophy of Science, 67(S3), S322–S331. https://doi.org/10.1086/392829

Gottlieb, G. (1965). Imprinting in relation to parental and species identification by avian neonates. Journal of Comparative and Physiological Psychology, 59, 345–356. https://doi.org/10.1037/h0022045

Gottlieb, G. (1978). Development of species identification in ducklings: IV. Change in species-specific perception caused by auditory deprivation. Journal of Comparative and Physiological Psychology, 92(3), 375–387. https://doi.org/10.1037/h0077473

Gottlieb, G. (1980). Development of species identification in ducklings: VI. Specific embryonic experience required to maintain species-typical perception in Peking ducklings. Journal of Comparative and Physiological Psychology, 94(4), 579–587. https://doi.org/10.1037/h0077691

Griffiths, P. E., & Hochman, A. (2015). Developmental systems theory. In eLS (pp. 1–7). Wiley. https://doi.org/10.1002/9780470015902.a0003452.pub2

Johnston, T. D. (2010). Developmental systems theory. In M. S. Blumberg, J. H. Freeman, & S. R. Robinson (Eds.), Oxford handbook of developmental behavioral neuroscience (pp. 12–29). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780195314731.013.0003

Kuo, Z. Y. (1930). The genesis of the cat’s responses to the rat. Journal of Comparative Psychology, 11(1), 1–36. https://doi.org/10.1037/h0075723

Lehrman, D. S. (1953). A critique of Konrad Lorenz’s theory of instinctive behavior. The Quarterly Review of Biology, 28(4), 337–363. https://doi.org/10.1086/399858

Lorenz, K. (1950). The comparative method in studying innate behaviour patterns. Symposia of the Society for Experimental Biology, 4, 221–268.

Michel, G. F. (2013). The role of developmental psychobiology in the unification of psychology. Review of General Psychology, 17(2), 210–215. https://doi.org/10.1037/a0032936

Oyama, S. (2000). The ontogeny of information: Developmental systems and evolution (2nd ed.). Duke University Press. https://doi.org/10.1515/9780822380665

Oyama, S., Griffiths, P. E., & Gray, R. D. (Eds.). (2001). Cycles of contingency: Developmental systems and evolution. MIT Press.

Turkheimer, E. (1998). Heritability and biological explanation. Psychological Review, 105(4), 782–791. https://doi.org/10.1037/0033-295x.105.4.782-791

Walsh, D. M., & Wiebe, K. (2020). The being of living beings: Foundationalist materialism versus hylomorphism. In A. S. Meincke & J. Dupré (Eds.), Biological identity: Perspectives from metaphysics and the philosophy of biology (pp. 107–127). Routledge. https://doi.org/10.4324/9781351066389-6

Witherington, D. C., Overton, W. F., Lickliter, R., Marshall, P. J., & Narvaez, D. (2018). Metatheory and the primacy of conceptual analysis in developmental science. Human Development, 61(3), 181–198. https://doi.org/10.1159/000490160

Explore Further

Greenberg, G. (2016). Zing-Yang Kuo (Part 1). The Behavioral Neuroscientist and Comparative Psychologist. https://www.apadivisions.org/division-6/publications/newsletters/neuroscientist/2016/12/kuo-part-1

A concise account of Kuo’s challenge to instinct explanations, including the cat-and-rat experiments and their developmental implications.

For a broader overview, see Griffiths and Hochman (2015), listed under Sources.