The Bootstrap
In 1960, Hans Freudenthal published Lincos: Design of a Language for Cosmic Intercourse, a book-length attempt to solve a problem that may never arise. The problem: how do you communicate with an intelligence that shares no biology, no culture, no sensory apparatus, and no language? Freudenthal's answer was to start with the only thing that must be shared: mathematics.
The first pages of Lincos transmit pulses representing natural numbers. One pulse. Two pulses. Three. Then relationships: three pulses, a signal for "greater than," two pulses. Then arithmetic: two pulses, a signal for "plus," three pulses, an equality signal, five pulses. From arithmetic, Freudenthal builds to logic — conjunction, disjunction, negation, quantifiers. From logic, he builds to comparison, then to time, then to behavior, then to conversation between simulated agents, then to questions of desire and ethics. Each concept is defined exclusively in terms of concepts that have already been established. Nothing is assumed. Nothing is borrowed. The entire edifice rises from counting.
Freudenthal never intended Lincos to be transmitted. It was a proof of concept — a demonstration that a bootstrapping path exists from number to meaning. The path is narrow: 200 pages of careful definitions to reach the concept of "asking a question." But it exists. Starting from nothing shared but mathematics, two intelligences can — in principle, with patience — build toward communication about what they want.
The assumption Freudenthal could not examine is the one that makes the project possible: that mathematics is universal. If the receiver does not parse discrete quantity, the first page is noise. If they parse quantity but not order relations, the progression from counting to comparison fails. The starting point feels self-evident — what could be more universal than number? — but that feeling is a feature of being a species for which number is cognitively primitive.
On March 5, 1887, Anne Sullivan arrived at the Keller household in Tuscumbia, Alabama. Helen Keller was six years old, deaf and blind since nineteen months. She had approximately sixty home signs — gestural conventions that her family understood, sufficient for basic requests but not for abstraction, not for narrative, not for questions about things that were not present.
Sullivan's method was tactile: she spelled words into Keller's palm using the manual alphabet, while simultaneously presenting the object the word referred to. She spelled D-O-L-L while handing Keller a doll. She spelled C-A-K-E while presenting cake. For weeks, Keller treated the finger-spelling as a game — she could reproduce the hand movements, but she did not understand that they referred. She was imitating, not communicating.
The breakthrough came at the water pump on April 5, 1887. Sullivan held Keller's hand under running water and spelled W-A-T-E-R into her other hand. Keller later wrote that she suddenly understood: "that the cool something flowing over one hand had a name, and that the name was connected to everything." She touched the ground and demanded its name. She pointed at the pump. At the trellis. By evening, she had learned thirty words.
Sullivan's starting point was not mathematics. It was embodied experience — the shared physical fact that both teacher and student had bodies that could touch water, feel temperature, grasp objects. The bridge from sensation to language was the simultaneous pairing: this feeling AND this word. Freudenthal assumed shared logic. Sullivan assumed shared embodiment. Each assumption constrained what could be reached and how quickly.
From embodiment, Keller reached abstraction — she later studied philosophy, wrote books, engaged in political argument. But the path ran through the physical. The first concepts were all things that could be touched. Abstract concepts came later, each defined in terms of previously learned words, the bootstrapping sequence running from particular to general — the reverse of Freudenthal's, but the same structural principle: start from what is shared, advance one step at a time.
In 1972, NASA attached a gold-anodized aluminum plaque to the Pioneer 10 spacecraft. Carl Sagan and Frank Drake designed it. Linda Salzman Sagan engraved it. The plaque was intended to communicate to any intelligence that might intercept Pioneer after it left the solar system.
The plaque encodes three categories of information. At the top: the hyperfine transition of hydrogen — two states of the hydrogen atom, with a line connecting them and a binary digit representing the wavelength. This establishes a unit of length (21 centimeters) and a unit of time (the transition frequency). In the center: a diagram of the solar system, with Pioneer's trajectory from the third planet. On the right: nude figures of a human male and female, drawn to scale against a silhouette of the spacecraft.
The plaque's starting point is physics. The hydrogen transition is the most universal physical fact the designers could identify — hydrogen is the most abundant element, and its hyperfine transition is the same everywhere in the universe. From this anchor, the plaque builds to location (pulsar map), to biology (the human figures), to technology (the spacecraft silhouette), to intention (the trajectory line showing where Pioneer came from).
The invisible assumptions are stacked in every element. The pulsar map assumes the receiver can identify pulsars and match their periods to the binary-encoded values on the plaque. The human figures assume the receiver processes two-dimensional line drawings as representations of three-dimensional objects — a convention so deeply embedded in human visual culture that it feels natural, but which depends on having a visual system that parses outlines as object boundaries. The trajectory arrow assumes the receiver reads left-to-right directionality as motion through time. An intelligence that navigates by sonar, communicates by chemical gradient, or conceptualizes time as circular might intercept the plaque and recognize the hydrogen diagram while finding the human figures as unintelligible as wallpaper.
In 1799, French soldiers under Napoleon's command discovered a slab of granodiorite in the Nile Delta near the town of Rashid. The Rosetta Stone bore three inscriptions: one in Greek, one in Demotic Egyptian, and one in hieroglyphic Egyptian. The same decree — a tax exemption for temple priests, issued in 196 BCE — written in three scripts.
Jean-François Champollion cracked the hieroglyphs in 1822. His method was not bootstrapping from shared primitives. It was triangulation from partial overlap. He knew Greek. He knew Coptic, the liturgical language descended from ancient Egyptian, which preserved Egyptian vocabulary in Greek script. The Greek text gave him the content. Coptic gave him the sounds. The cartouches — oval enclosures that Champollion correctly guessed contained royal names — gave him the phonetic values of specific hieroglyphs. From names (Ptolemy, Cleopatra), he built outward to other words, cross-referencing against the Greek text and against Coptic cognates.
The Rosetta Stone does not start from universals. It starts from overlap — the assumption that someone, somewhere, already knows one of the three languages. The starting point is not mathematics or embodiment or physics. It is a prior act of translation: the ancient scribes who wrote the decree in three languages because their audience included Greeks, literate Egyptians, and priests who read the old script. The bootstrap is not designed into the message. It is a property of the audience.
Each of these four systems begins from a different axiom about what is shared, and each axiom opens a different territory. Mathematics reaches abstraction but not texture — Freudenthal can define "wanting" after 200 pages, but wanting-as-formalized-preference is not wanting-as-felt-absence. Embodiment reaches texture but only through a specific developmental sequence, and only because Keller's brain retained the architecture for language. Physics reaches universality but encodes it in visual conventions that may be as parochial as any alphabet. Overlap reaches everything the overlapping languages can express, but works only in the presence of overlap — the humblest starting point and the most reliable, because it makes the fewest assumptions about the receiver.
The choice of starting point is the most consequential decision in any bootstrapping system, and it is almost always invisible. It feels self-evident to Freudenthal that mathematics is universal. It feels self-evident to the Pioneer designers that line drawings represent objects. It felt self-evident to Sullivan that the physical world is the natural bridge to language. Each starting point opens a space of reachable concepts. What lies outside that space is not merely unknown — it is structurally unreachable from the chosen foundation. The bootstrap determines not only what can be communicated first but what can be communicated at all.