What Is Life? Key Insights on Purpose, Agency, and Artificial Life

The discussion revolves around the fundamental question of "what is life and how does it matter," bringing together perspectives from science and philosophy. The speakers, Mel, Phil, and Nick, explore the historical evolution of this question, current scientific advancements, and the ethical implications of new technologies like artificial life and AI.

The Nature of Purpose in Life

Mel introduces the topic by challenging Richard Dawkins' 1991 Christmas lecture assertion that "nature exists without purpose." While acknowledging the evolutionary basis of human brains and their role in experiencing purpose, Mel proposes that life itself might be inherently purposive, rooted in "embodied agency."

To illustrate this, Mel conducts a simple experiment: 1. Hold your thumb out and move your head side to side. Your thumb remains in focus. This is due to rapid, unconscious corrections by your inner ear and eyes, demonstrating a whole-body action not consciously willed. 2. Hold your thumb out and move your arm side to side at the same rate. Your thumb becomes blurry. This highlights the difference when the entire organism isn't recruited in the same integrated way.

This experiment emphasizes that the "wholeism" of the body, where the entire organism is recruited in actions, might be fundamental to what it means to be alive.

The Elusive Definition of Life

The history of science is filled with attempts to define life, often without consensus. Michael Levin's efforts to gather definitions from numerous scientists and even AI failed to yield a definitive answer. This difficulty stems from shifting boundaries of what is considered alive: * A hummingbird is clearly alive. * A single-celled methanogen is alive. * A slime mold (e.g., Physarum polycephalum) exhibits intelligent behavior and is alive. * A coral reef, a complex ecosystem of symbiotic organisms, raises questions about whether the ecosystem itself is alive. * James Lovelock's Gaia theory even proposed the Earth as a potentially living entity.

Artificial Life and Its Implications

Phil introduces a recent attempt to create "lifelike" systems from scratch. Kate Adamala's work at the University of Minnesota created "spud cells" with the following characteristics: * Structure: Boundaries made of lipid membranes, similar to cell membranes. * Components: Contain synthetic genomes (36 genes), enzymes, and ribosomes (protein-making molecular machines). * Lifelike activities: Can replicate their genomes, divide (with some external help), and undergo competition and a form of natural selection. Variants with genes for faster growth become more successful.

However, these spud cells are not considered alive because: * They are entirely dependent on a rich chemical "soup" and external support. * They cannot build their own ribosomes. * They struggle to control cell division effectively, leading to uneven gene distribution in daughter cells. * They "run out of steam" after a few replication rounds.

This experiment highlights the complexity of true aliveness and the gap between lifelike systems and actual life.

Mel then connects this to the rise of Artificial Intelligence (AI) and its impact on human perception: * Increased belief in AI intelligence: Just three exposures to AI can increase belief in its intelligence (Clara Colomb & Jonathan Birch). * Trust in anthropomorphic AI: More anthropomorphic AI models lead to greater trust in their answers, even if incorrect (Michelle Conn). * Reduced reliance on humans: Five minutes a day talking to an "agentic" AI about personal problems can significantly reduce the preference for turning to humans (Yaoi Xi).

These findings underscore the ethical and moral urgency of clarifying what it means to be alive, intelligent, and to possess agency and sentience in an age of artificial life and intelligence.

Historical Perspectives on Life

Phil provides a historical overview of how life has been understood:

Ancient Times (Aristotle)

• Life was seen as a fundamental organizing principle of the cosmos, not something to be explained by metaphors, but rather the metaphor itself.
• Aristotle believed living things possessed a "psyche" (often mistranslated as soul), which was an innate capacity for action, inseparable from the body.
• The psyche gave organisms capacities for growth, self-nourishment, movement, and perception.

17th Century (Mechanism)

• With Isaac Newton's laws of motion, natural philosophers began to view nature as bodies interacting through forces.
• René Descartes proposed the human body as a mechanism (pumps, bellows, levers), but insisted on a divinely granted, independent soul to avoid heresy.
• Julien Offray de La Mettrie, in the 18th century, challenged this, suggesting life was an innate property of the body, calling it a "self-winding machine." His book "Man a Machine" presented humans as complex machines, differing from animals only in complexity.

18th-19th Centuries (Chemistry and Vitalism)

• Focus shifted to chemistry, with Antoine Lavoisier noting carbon's presence in both living things and non-living diamonds.
• Georges-Louis Leclerc, Comte de Buffon, proposed "matière vivante" (living material) composed of active, indestructible molecules, suggesting a "vital force."
• Félix Dujardin thought he isolated this living matter, calling it "sarcode" (later protoplasm). Thomas Henry Huxley also claimed to find it in deep-sea sediments, though it was later found to be a chemical artifact.
• This "vital force" idea was criticized as tautological.
• Jöns Jacob Berzelius in 1812 argued against a vital force, stating that "the power to live... is the result of the mutual operation of the instruments and the rudiments on one another," emphasizing organization and interaction.
• The discovery of cells and their internal organization (mitochondria, chromosomes) shifted the focus to understanding these structures.

20th Century (Molecular Biology and Gene-Centrism)

• The mission became to understand life at the molecular level.
• François Jacob (Nobel laureate) stated, "The aim of modern biology is to interpret the properties of the organism by the structure of its constituent molecules."
• DNA, with its information-encoding capacity (discovered by Watson and Crick), became central.
• Ernst Mayr's concept of a "genetic program" emerged, suggesting all organismal activity is influenced by genes.
• Jacob famously declared in 1970, "Biologists no longer study life today. Biologists study the molecules of life, but life itself... who needs it?"
• Richard Dawkins' "selfish gene" theory (50 years old this year) posited genes as agents, with organisms as mere "passive vehicles."
• This gene-centric view, according to the speakers, has led to life itself "vanishing entirely" from biological discourse.

The Problem with Current Definitions of Life

Nick critiques NASA's working definition of life: "a self-sustained chemical system capable of undergoing Darwinian evolution." * "Self-sustained": Nick argues organisms are not self-sustained but are sustained by their environment. * "Capable of undergoing Darwinian evolution": Taken strictly, this implies a single rabbit isn't alive, only a pair capable of reproduction.

He also challenges the idea that life requires "liquid water and organic polymers such as nucleic acids and proteins," as a dead whale contains all these components but is clearly not alive. The missing element is energy flow.

Energy Flow and the Origin of Life

Nick's obsession is understanding how energy works in biology. He highlights Peter Mitchell's work on cell respiration: * Mitchell realized that life isn't just "chemicals in a bag" but involves the structure of cells and electricity. * In 1957, Mitchell stated, "I cannot consider the organism without its environment... the two may be regarded as equivalent phases between which dynamic contact is maintained by the membranes that separate and link them." This "conflation" of organism and environment is crucial for understanding life's origins.

Nick explains the process of respiration in mitochondria (or bacteria): * Food is broken down, releasing carbon dioxide and hydrogen. * Hydrogen (electrons and protons) is "burned" with oxygen. * Electrons flow to oxygen, powering the extrusion of protons across a membrane. * This creates an electrical charge across the membrane (e.g., 150-200 millivolts across 5 nanometers, equivalent to 30 million volts per meter – like a lightning bolt). * This voltage is the "system" that separates the cell from the world.

The question then becomes: how did such a sophisticated, interwoven system (proteins pumping protons, ATP synthase as a rotating nanomachine, impermeable membranes) arise?

Nick proposes a model for the origin of life in hydrothermal vents: 1. Natural Proton Gradient: Hydrothermal vents have acidic ocean waters and alkaline hydrothermal fluids, creating a natural proton concentration difference across porous rock structures (micrometers in size, like cells). This provides the "voltage" naturally. 2. Metabolism: This natural gradient can drive chemical reactions, such as converting hydrogen and carbon dioxide into the molecules of life. Nick argues that metabolism is not invented by genes but is "thermodynamically favored chemistry" that would spontaneously occur under the right conditions. 3. Membrane Formation: Fatty acids spontaneously form membranes around aqueous spaces in water, a thermodynamically favored process that increases entropy. 4. Genes Emerge Later: Genes arise later, influencing existing metabolism and the body, rather than dictating everything from the start. This suggests that "metabolism and the body... came before genes."

Reclaiming the Organism and the Body

Mel argues that the most remarkable aspect of life – its purposeful action – has been sidelined. Aristotle's focus on "what animates life" and the necessity of "purposiveness" is relevant. He believed in a "holism" where matter, form, and directionality (metabolism/energy) come together in an integrated, embodied entity.

Mel criticizes the historical "dualism" that separates the animating principle from the body: * Plato's immortal soul. * The "immortal gene" (Dawkins' original title for The Selfish Gene). * The mind/brain as the sole driving principle.

This dualism "loses this cleanest and clearest explanation for that remarkable purposeful quality of life because you don't have the organism anymore."

Challenges in Identifying Agency

Nick highlights the difficulty in identifying agency: * Over-attribution: Humans tend to "misattribute agency" where none exists, like seeing a "monster" in a bubble interacting with fatty acid structures (which are just self-organizing due to chemistry). This is akin to the "Pac-Man effect." * Under-attribution: Conversely, we fail to see agency where it clearly exists, such as in a Caledonian crow using a tool or a slime mold efficiently mapping a subway system. We dismiss them as "machines running on their genes."

Towards a Theory of Biological Agency

Phil discusses the controversial idea of biological agency as an attempt to explain the fundamental difference between living organisms and non-living things (rocks, hurricanes). While goal-directedness is acknowledged by many biologists, the source of these goals is debated.

• Genetic Program View: Organisms are "automata directed by their genes," with natural selection favoring genes that promote survival. Phil argues this is insufficient, as organisms need to innovate and improvise in novel situations, suggesting evolution would favor agents.
• Definition of Agency: Most definitions boil down to "the ability to manipulate the entity itself and its environment in order to attain some goals," which the entity itself determines.
• Critique of Agency: Some philosophers (James DeFrisco, Richard Gorn) question the utility of "agency" if it doesn't do explanatory work beyond describing molecular actions.
• Naturalizing Agency: Phil, aligning with neuroscientists Kevin Mitchell and Henry Potter, believes agency can and should be "naturalized" into a rigorous scientific idea. This involves identifying properties needed for a genuine agent:
  • A boundary separating it from the environment.
  • Internal complexity for different internal states (genes lack this).
  • Capacity to make predictions about consequences of actions (seen even in single-celled organisms).

A proper theory of agency could help evaluate the "real agency" of AI systems and inform ethical considerations.

The Body and the Genome

Mel emphasizes the historical sidelining of the body and the brain-centric view of agency. Antonio Damasio's quote, "If ensuring the survival of the body proper is what the brain first evolved for, then when minded brains first appeared, they began by minding the body," highlights the body's primacy. Many organisms exhibiting remarkable agency (bacteria, amoeba, hydra, tardigrades, oak saplings) lack brains.

Mel suggests resistance to a body-centered view stems from the body's mortality, messiness, and vulnerability. She criticizes the persistent "body as machine" metaphor and Dawkins' statement: "We are machines built by DNA whose purpose is to make more copies of the same DNA."

Phil counters this gene-centric view: * Information Flow: The traditional "bottom-up" view (genes -> proteins -> phenotype) is not supported by modern molecular biology. Information flows "up and down and sideways and always." * Cellular Differentiation: All cells in the body have the same genome but perform different jobs (liver, neuron, muscle). This is not solely due to genetic programming but to information received from the environment and other cells, which "turn on and off different suites of genes." * Open Informational Systems: Only open informational systems can become agents. * Genome as Resource: Phil suggests ditching the "program" notion for the genome and viewing it as a "source of molecular resources" that living cells use to develop morphology.

Scaling Agency and Ethical Considerations

The speakers then explore how agency scales: * Single-celled to Multicellular: How does agency transition from a single-celled organism to a multicellular one (e.g., a human)? Are all cells agents, or is there a different agent at the multicellular level? * Colonies and Usocial Organisms: What happens in a colony or usocial organism (like ants) where individuals are interdependent, almost forming a "distributed body"? * Symbiosis: In tight symbiotic relationships, where does the agent begin?

Nick discusses the origin of eukaryotic cells (which make up plants, animals, and amoeba): * Eukaryotic cells arose only once in 4 billion years of life. * They always contain mitochondria, which were once free-living bacteria. * This involved a "shift in the level of agency" and a "reorganization of the self." The incorporated bacteria had to lose their individual agendas, becoming integrated into a composite self. * This allowed eukaryotes to scale up their genomes and regulatory complexity. * Multicellularity (30-40 separate origins) and complex organisms (plants, animals) represent further shifts in agency to composite selves.

Nick highlights gaps in scientific understanding: we don't fully know how mitochondrial function and plasma membrane activity are integrated at the cellular level.

Phil then discusses Michael Levin's work, which further destabilizes the genetic program notion: * Developmental Plasticity: The outcome of a genome isn't fixed. A butterfly genome can produce a caterpillar or a butterfly. * Xenobots: Levin's frog cells, when clumped, can form "xenobots" that behave like organisms, swimming and moving, raising questions about whether they are a new type of living entity. * Organoids: Human cell cultures can form organoids (e.g., brain organoids from skin cells) that resemble embryonic brains or even entire embryos without eggs or sperm. This shows multiple possible developmental outcomes from the same genome. * Genome as Generative Algorithm: Kevin Mitchell suggests viewing the genome as a "generative algorithm" rather than a program. Phil prefers seeing it as a "source of molecular resources."

Finally, Phil touches on the ethical implications of brain organoids: * As organoids become more brain-like, questions arise about their capacity for sentience or consciousness. * If they possess these, what ethical obligations do we owe them?

The speakers conclude by emphasizing that this is a rapidly evolving field where old models are breaking down, and new insights are challenging our fundamental definitions of life, agency, and intelligence, with significant ethical consequences. They invite questions from the audience, promising to share their own (potentially differing) definitions of life at the end.