How Landscape Metaphors Shape Science
Why a 90-Year-Old Visual Metaphor Still Dominates How We Think About Change
Imagine evolution as a rugged mountain range, with peaks representing optimal designs and valleys of failure. This compelling image, known as the fitness landscape, has guided scientific thought for nearly a century, creating an intuitive bridge between abstract theory and tangible reality 16. First conceived by geneticist Sewall Wright in 1932, this powerful metaphor has transcended its biological origins to influence how researchers understand everything from cultural trends to technological innovation 249.
Recent discoveries, however, are revealing startling flaws in this classic model. Scientists examining everything from genetic data to religious practices are finding that the evolutionary landscape might be less like a range of climbable mountains and more like dangerous Swiss cheese—seemingly flat but riddled with invisible trapdoors that can spell instant extinction for species wandering in the wrong direction 1. This paradigm shift is forcing a fundamental reexamination of how evolution actually works.
Fitness Peak
Fitness Valley
Evolutionary "Hole"
Sewall Wright's original fitness landscape presented a revolutionary way to visualize evolution. In this model, population of organisms evolve toward higher fitness peaks (adaptation) while avoiding lower fitness valleys 6. The "peaks" represented optimal genetic combinations that maximized survival and reproduction, while "valleys" represented disadvantageous combinations 17.
This metaphor successfully unified disparate evolutionary phenomena under one intuitive framework and became central to what biologists call the Modern Synthesis—the reconciliation of Mendelian genetics with Darwinian natural selection 67.
The traditional landscape model began showing cracks when scientists considered quantitative traits—characteristics like flight capability or complex physiological processes that depend on multiple components working together simultaneously 1.
This interdependence suggests that evolutionary progress isn't about steady climbing, but about avoiding catastrophic failures. As Sergey Gavrilets noted, the landscape might actually be "holey"—mostly flat with average fitness, but punctuated by holes representing inviable combinations that lead to extinction 1.
Another limitation emerged regarding environmental stability. Wright's original model assumed relatively stable landscapes, but real environments constantly change. A peak适应 for a savannah environment becomes a valley during desertification 1. This insight led Mustonen and Lässig in 2009 to propose the "fitness seascape" metaphor instead, emphasizing how evolutionary environments fluctuate like ocean currents 1.
| Metaphor | Time Period | Key Characteristics | Primary Limitations |
|---|---|---|---|
| Fitness Landscape | 1932-present | Stable peaks and valleys; gradual climbing | Assumes environmental stability; oversimplifies trait interdependence |
| Rugged Landscape | 1990s-present | Multiple local optima; harder navigation | Still assumes Gaussian distributions and climbable paths |
| Fitness Seascape | 2009-present | Dynamic, changing environments | Doesn't address fundamental issues with trait interdependence |
| Holey Landscape | 1997-present | Flat with "trapdoors"; inviable combinations | Counters intuitive notion of evolutionary progress |
A groundbreaking 2024 study published in PNAS, "Drift on holey landscapes as a dominant evolutionary process," put the holey landscape hypothesis to rigorous testing 1. The research team from universities in North Dakota, California, and Paris took these steps:
The team gathered genetic variation data from sixty different species, including diverse animals and plants, creating a comprehensive dataset of evolutionary trajectories 1.
They simulated evolutionary pathways using both traditional Gaussian landscape models and holey landscape models, comparing how populations evolved under each scenario 1.
Researchers compared empirical genetic patterns from natural populations against predictions generated by both landscape models, looking for which model better explained observed reality 1.
Unlike traditional one or two-trait landscapes, the team analyzed how multiple interdependent traits behave simultaneously—what they called "high-dimensional" analysis 1.
The findings challenged decades of evolutionary biology dogma. The patterns observed in natural populations consistently aligned with holey landscape predictions rather than traditional models 1.
| Analysis Type | Traditional Landscape Prediction | Holey Landscape Prediction | Actual Observation |
|---|---|---|---|
| Trait distribution | Gaussian curves with smooth transitions | Flat averages with discontinuous "holes" | Consistent with holey landscape |
| Evolutionary paths | Gradual climbing between peaks | Drift across flat regions until finding viable combinations | Supported holey landscape model |
| Multi-trait integration | Additive, independent effects | Holistic thresholds requiring simultaneous coordination | Contradicted traditional models |
| Fitness optimization | Clear peaks and valleys | Peaks "average out" when considering multiple traits | Favored holey landscape view |
The researchers concluded that "our understanding of how evolution has shaped phenotypes remains incomplete" and that "simple standard evolutionary models are not consistent with available data for quantitative data" 1. This suggests that evolution often proceeds not by climbing toward optimal peaks, but by drifting across flat regions while avoiding disastrous holes—like navigating a minefield rather than climbing a mountain 1.
Comparison of how well different evolutionary models explain observed genetic patterns in natural populations.
The landscape metaphor has proven remarkably versatile, extending far beyond its biological origins to help understand human culture and technology.
Researchers from Carnegie Mellon University and the Santa Fe Institute applied landscape mathematics to 407 world religions, treating belief systems as existing on a cultural landscape 29. Using Bridges-2 supercomputing power, they analyzed how religions evolve, persist, or die out 9.
| Religion Type | Landscape Position | Stability Characteristics | Example |
|---|---|---|---|
| State-endorsed religions | High, stable peaks | Resistant to change; persistent over time | Historical state religions |
| Evangelical religions | "Floodplain" regions | Stable yet adaptable to change | Modern evangelical movements |
| Mystery religions | Intermediate elevations | Moderate stability with some flexibility | Ancient mystery cults |
| Extreme traditions | Unstable cliffs | Rapid collapse; don't persist | Human sacrifice practices |
The research revealed why some religious practices persist while others disappear: "State-endorsed religions experienced stability. Evangelical religions, non-state-sponsored religions and mystery religions each had unique stability—existing on a 'floodplain' that offers stability and the ability to change" 2.
Frank Geels and others have used multi-level perspective (MLP) landscapes to understand technological transitions, visualizing society as having "landscape levels" that are hard to change, "regime levels" with established practices, and "niche levels" where innovations emerge 8. This approach helps explain why some technologies succeed while others fail, depending on their position in the socio-technical landscape.
Macro-level trends, slow to change
Established practices and rules
Innovation and experimentation
The landscape metaphor approach requires specific conceptual and methodological tools across different fields:
| Research Tool | Field of Use | Function | Example Application |
|---|---|---|---|
| High-dimensional statistical analysis | Evolutionary biology | Models multiple interdependent traits simultaneously | Analyzing how flight-related traits co-evolve 1 |
| Unrestricted Boltzmann machines | Cultural evolution | Corrects for incomplete and biased historical data | Reconstructing unknown aspects of poorly documented religions 9 |
| Bayesian logic frameworks | Cultural evolution | Determines likely relationships between factors despite missing data | Estimating whether undocumented religions believed in moralizing gods 9 |
| Multi-level perspective (MLP) | Technological studies | Analyzes interactions between landscape, regime, and niche levels | Understanding renewable energy adoption barriers 8 |
| Holey landscape modeling | Evolutionary biology | Simulates evolution on flat landscapes with inviable "holes" | Testing why some genetic combinations never appear in nature 1 |
Advanced statistical methods are essential for analyzing complex evolutionary patterns across multiple dimensions and identifying subtle landscape features.
Simulation models allow researchers to test different landscape configurations and evolutionary scenarios that would be impossible to observe directly.
The landscape metaphor continues to be an astonishingly productive scientific tool nearly a century after its introduction, despite its known limitations. As philosopher Stefan Petkov noted, the fitness landscape is "dead but not gone"—scientifically problematic yet indispensable as a conceptual framework 67.
Recent discoveries about holey landscapes don't invalidate the metaphor so much as transform it. Rather than climbing ever higher, evolution often involves navigating flat plains while avoiding catastrophic trapdoors 1. This revised understanding better accounts for phenomena like irreducible complexity, where multiple components must emerge simultaneously for a system to function at all 1.
The metaphor's true power lies in its flexibility across disciplines—from genes to memes, researchers continue finding value in visualizing complex systems as navigable terrains. As science develops more sophisticated models, the landscape metaphor evolves with it, demonstrating that how we visualize scientific concepts fundamentally shapes what we can discover about the natural world.
The challenge moving forward is developing what the PNAS researchers called "clear alternative explanations besides a simple null hypothesis of drift with no selection" 1. As we refine our mental maps of evolutionary processes, we inevitably refine our understanding of life's incredible diversity—and our place within it.
References will be added here manually.