StarTalk Cosmic Queries: Terence Tao on Pure vs Applied Mathematics and the Unreasonable Effectiveness of Math

In this StarTalk Cosmic Queries episode, host Neil deGrasse Tyson talks with Terence Tao about the differences between pure and applied mathematics, the use of toy models like the spherical cow, and why mathematics so often captures the behavior of the real world. The discussion covers unsolved problems such as the Collatz conjecture, the power of crowdsourcing in mathematical research, and the interplay between mathematical theory and practical science. The hosts also explore base systems, chaos, Gaussian distributions, and how mathematics informs physics, computation, and education. The conversation demonstrates how collaboration across disciplines can accelerate discovery and why partial progress is still valuable in mathematics.

Overview and Episode Context

The StarTalk Cosmic Queries episode centers on mathematics, featuring Terence Tao as the guest and Neil deGrasse Tyson as host. The conversation begins with a nod to Vladimir Arnold’s idea that mathematics is the part of science where experiments are cheap, setting the tone for a discussion about the role of mathematical reasoning in science. The hosts explore why our current mathematical framework explains a large portion of the universe, yet where and why our models begin to fail when faced with quantum space‑time, singularities, or the extreme regimes of physics.

Introduction to Pure and Applied Mathematics

The discussion clarifies the core distinction between pure mathematics, which pursues patterns and structures for their own sake, and applied mathematics, which seeks mathematical descriptions that are useful to scientists and engineers. Tao emphasizes that modern science is highly interdisciplinary, making collaboration across fields essential. The dialogue stresses that mathematics acts as a tool for compressing complex data and physics, enabling scientists to forecast, simulate, and design experiments with greater efficiency and efficacy.

They discuss how toy models, such as spheres or frictionless bodies, often provide the simplest path to understanding a phenomenon. The spherical cow analogy is introduced to illustrate how simplifying assumptions help isolate the dominant effects while bounding the complexity of a model. The contrast with real world systems is highlighted: comfortable abstractions can reveal fundamental limits and guide subsequent refinements that gradually approach reality without becoming intractable.

Throughout, Tao explains that the flexibility of mathematics allows researchers to alter hypotheses and explore negative results, proving to be a powerful diagnostic tool that reveals where assumptions fail and what additional ingredients are needed.

Interdisciplinary Collaboration and Center‑Stage for Applied Math

The IPA Institute for Pure and Applied Mathematics and Tao’s role as director of special projects are introduced as a platform for bringing together pure mathematicians, applied scientists, engineers, and industry professionals. The format enables a cross pollination of ideas long before practical technologies emerge. Tao shares anecdotes about how early interactions with electronic engineers and statisticians led to improvements in MRI technology, which are now widely used in medical imaging. The dialogue emphasizes that collaboration is often the key to turning abstract mathematics into real world impact.

Pure versus Applied: Definitions, Boundaries, and the Value of Toy Models

The hosts define pure mathematics as curiosity driven exploration of patterns in numbers and shapes, while applied mathematics focuses on the mathematical structures most relevant to real world phenomena. They discuss how mathematicians can act like “toy model designers,” choosing the simplest version of a problem in order to extract the essential features. The sphere analogy is revisited: by starting with a spherical cow, researchers can bound the problem and later incorporate more realistic details, ensuring progress while maintaining tractability.

The conversation draws a distinction between mathematical modeling in physics and the demands of other sciences. Tao notes that in mathematics, hypotheses can be altered, and errors can be explored without endangering people, whereas in medicine or engineering, failures can be catastrophic. This difference underlines why math has a unique capacity for risk free exploration, enabling rapid iteration and learning from mistakes.

Unsolved Problems, Crowdsourcing, and the Nature of Progress

A major portion of the dialogue centers on unsolved problems in mathematics and the value of approaching them from multiple angles. The Collatz conjecture is presented as a deceptively simple process with an open question about whether all numbers eventually reach one under a simple iteration. Tao and Tyson discuss the role of computational experimentation and large scale testing in shaping conjectures while acknowledging that proofs are required for definitive resolution. The idea of crowdsourcing problems, particularly Erdős problems, is explored as an effective, decentralized approach to generating ideas and partial breakthroughs.

The pair describe a recent instance where a problem from Erdős's problem database (1026) was advanced through a collaborative process with about 56 participants. They use this example to illustrate a broader point: modern mathematics benefits from diverse contributions, and a formal paper may not be the only or final path to a breakthrough. The community can progress through iterative, shared exploration that builds on prior insights, even if the immediate solution remains partial or provisional.

Collatz Conjecture as a Case Study in Chaos

The Collatz problem is revisited to illustrate how extremely simple rules can give rise to unexpectedly complex and chaotic behavior when iterated. Tao and Tyson explain how, despite the problem’s simplicity, the long term behavior of sequences can be highly unpredictable. They connect this to a broader phenomenon in chaos theory: simple dynamical rules can generate intricate and seemingly random patterns over time. The discussion emphasizes that solving such problems often requires a blend of rigorous theory and empirical evidence, along with an openness to unexpected directions and new ideas.

Gaussian Distributions and Universality in Nature

The episode turns to the ubiquity of Gaussian distributions and other universal patterns across many scientific disciplines. The central limit theorem is highlighted as a foundational explanation for why bell curve shapes pop up in diverse contexts, from human heights to measurement errors. Tao and Tyson discuss the idea that some statistical laws are robust and appear due to underlying probabilistic principles, rather than specifics of a single system. They also touch on other distributions observed in nature, noting that while some universal patterns are well understood, others remain less explained and are the subject of ongoing mathematical inquiry.

Base Systems, Computation, and Cognitive Considerations

A thought provoking question about using a base other than ten leads to a discussion of how counting systems, and the binary base used in computing, shape mathematical thinking and computational practice. They acknowledge historical bases such as base 60 in ancient Babylonian mathematics and base 12 remnants in language and culture, as well as base 20 counting in some contexts. The hosts conclude that while base 2 is optimal for computation, base 10 remains efficient and intuitive for everyday tasks, and that changing bases would not fundamentally alter the potential for mathematical discovery.

Beyond Euclidean Geometry and the Quantum Gravity Frontier

Reflecting on historical mathematics, the conversation addresses non Euclidean geometries and their deep connections to physics, notably general relativity. They discuss how the geometric language of curved space became essential to Einstein’s theory and how mathematicians like Riemann provided the tools that enabled relativity. The discussion then turns to the pursuit of a quantum theory of gravity, noting that current frameworks may require new mathematical structures beyond non Euclidean geometry to describe space time at the smallest scales and the largest energies. The host and guest acknowledge Ed Witten’s involvement with string theory while yet remain skeptical that the current mathematical formulation fully matches experimental reality.

Future of Mathematical Discovery and the Human Element

The concluding sections emphasize that mathematics, while deeply abstract, is ultimately a human endeavor guided by intuition, creativity, and collaboration. Tao discusses the role of negative results in clarifying the boundaries of what can be proven, and the value of speculative progress when it opens new lines of inquiry. The dialogue also touches on the portrayal of mathematicians in popular culture, the emotional engagement required to teach and popularize math, and the importance of motivating educators to connect with students through enthusiasm and clear communication. The conversation closes with a shared sense that math is a continually evolving language for describing the universe, with new ideas likely to emerge from interdisciplinary collaborations and novel computational tools.

Implications for Research and Education

Overall the episode highlights how mathematical thinking informs not only theoretical insights but also practical technology and scientific methods. It showcases a culture that values curiosity, cross disciplinary collaboration, and the iterative nature of problem solving. The discussion implicitly encourages budding mathematicians and scientists to embrace both abstract reasoning and empirical exploration, recognizing that progress often comes in incremental steps and through the work of many contributors across a spectrum of disciplines.