The Symbolic Bet

Charniak Ch. 1–2 · Dartmouth, symbols, and the birth of classical AI

HON 201 · Week 8

Jane and the Piggy Bank

Jane was invited to Jack’s birthday party. She wondered if he would like a kite. She went to her room and shook her piggy bank. It made no sound.

Why did Jane shake the piggy bank?

With a partner: list every fact you needed to answer that.

Cold open, about 8 min. Let pairs work for 4 minutes, then collect a list on the board and count the items. Keep the list up; the game uses it.

1956: The Promise

A Summer at Dartmouth

“The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.”

— Proposal for the Dartmouth Summer Research Project on AI (1955) McCarthy, Minsky, Rochester, Shannon

“We think that a significant advance can be made … if a carefully selected group of scientists work on it together for a summer.”

Ask the class to put it in Kuhn’s terms: is this a theory, a paradigm, or something earlier? It’s a promise of a paradigm, close to Kuhn Ch. II’s pre-paradigm schools. There are no exemplars yet, only a conjecture and a timeline.

What They Had to Work With

General Electric GE-235 (1964 brochure)

Dartmouth’s 1964 time-sharing system, where BASIC was born, ran on the GE-235’s sibling, the GE-225. Ask students what they notice in the photo: rooms, desks, ties, paper.

The Machine, by the Numbers

	GE-235 (1964)	Your phone
Memory	~40 KB	~8 GB
Speed	6 µs / cycle	< 1 ns

Your phone has roughly 200,000× the memory of the biggest GE-235.

“The Balanced Computer”

EAI 640 (1966): “a price which starts under $30,000”, about ten times that in today’s dollars

The EAI 640 has 16-bit words, up to 32,768 words of memory, and a teletype. Note how the brochures sell the future: computers in a garden, beside a fountain. Ask: what is being promised here, and to whom?

The Bet: Thinking = Symbol Manipulation

The Physical Symbol System Hypothesis

“A physical symbol system has the necessary and sufficient means for general intelligent action.”

— Allen Newell & Herbert Simon (1976)

• Necessary: anything intelligent is manipulating symbols.
• Sufficient: the right symbol manipulation is intelligence. Nothing else is needed.

What a Symbol Means

(ISA   PIGGY-BANK  CONTAINER)
(HOLDS PIGGY-BANK  COINS)
(IF (AND (SHAKE ?X) (EMPTY ?X)) (NO-SOUND ?X))

To the machine, PIGGY-BANK means only what the rules say about it.

There is no picture of a pig, no memory of saving up, no sound.

So every fact on your list has to be written down by someone.

LISP (McCarthy, 1958) was the field’s language. Lists of symbols could hold both programs and data. This is the bridge to the game: they’re about to be the someone.

Early Wins

Look What Symbols Can Do

Year	Program	Feat
1956	Logic Theorist (Newell, Shaw & Simon)	proves theorems from Principia Mathematica
1959	Samuel’s checkers player	learns to beat its creator
1961	SAINT (Slagle)	does freshman calculus integrals
1964	STUDENT (Bobrow)	solves algebra word problems
1966	ELIZA (Weizenbaum)	“therapist” chat via pattern-matching
1970	SHRDLU (Winograd)	converses about a world of blocks

Notice the pattern: logic, games, algebra, and calculus, all formal domains.

SAINT

link

• p. 11: Leibniz
• p. 12: Goedel’s incompleteness theorem (bottom: “The Theorems show that the machine cannot do everything in mathematics. However, they can do a greal deal.”
• p. 13: Definition
• p. 58: Heuristic flowchart

John McCarthy: The Pioneer

• Background: Born in Boston to activist immigrant parents; taught himself calculus as a teenager
• AI Pioneer: Coined the term “Artificial Intelligence” (1955) and invented LISP (1958)

John McCarthy: The Vision

• Optimism: Firmly believed mathematical logic could formalize all human common sense and reasoning
• Outlook: Lifelong technological optimist; deeply committed to human progress through science and computing

Marvin Minsky: The MIT Lab & Perceptrons

• Co-founder of MIT AI Lab: Pioneered early AI, robotics, and cognitive science with McCarthy
• Perceptrons (1969): Influential critique co-authored with Seymour Papert analyzing limitations of single-layer neural networks (ushering in a shift toward symbolic AI)

Marvin Minsky: The Society of Mind

• The Society of Mind (1986): …suggested that intelligence emerges not from one system but from the interactions of numerous simple components, or “agents.”

Allen Newell: Human Cognition & AI

• Core Mission: Driven fundamentally by a desire to understand and model how humans actually think and solve problems
• Logic Theorist (1956): Co-created with Herbert Simon and Cliff Shaw; proved theorems from Principia Mathematica and is widely considered the first working AI program
• (Aside: Studied mathematics as a graduate student at Princeton at the same time as John Nash)

Herbert Simon: General Problem Solver

• Polymath Pioneer: Nobel laureate in Economics and Turing Award winner who bridged economics, psychology, and computer science
• General Problem Solver (GPS, 1959): Co-developed with Newell and Shaw; likely the first system designed to separate the general problem-solving engine from the domain-specific rules of the problem itself
• Bounded Rationality: Argued that human decision-making relies on heuristics and satisficing rather than pure optimization

The Paradox

“It is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility.”

— Hans Moravec (1988)

Why did it make sense to start with chess and calculus instead of children’s stories?

Moravec wrote this later, but the assumption behind it was there from the start: what’s hard for us must be hard for machines. The piggy bank shows the reverse.

The Culture

Optimism

“Machines will be capable, within twenty years, of doing any work a man can do.”

— Herbert Simon (1965)

“Within a generation … the problem of creating ‘artificial intelligence’ will substantially be solved.”

— Marvin Minsky (1967)

Enter Charniak (1967)

• Grad student at MIT’s AI Lab
• Thesis (1972): children’s stories
• Chosen because it looked easy

He’ll end up on both sides of the coming revolution.

Through Kuhn’s Eyes

A young community with:

• a shared commitment
• tools (LISP, logic, search)
• exemplars (SHRDLU …)
• solvable-looking puzzles

Is this a paradigm yet? Or still a promise?

Now: Program Jane

Groups of 3–4 take the roles:

• Rule-writer: writes the fact and rule cards
• Computer: uses only what’s on the cards
• Skeptic: counts and finds what you forgot

Start from your list of facts. You have 8 minutes.