Skip to content

A history reading path

Computing and artificial intelligence in history

Curated by HistorIQly · Updated

What could a machine calculate, what would count as intelligent behavior, and what did researchers propose to investigate? Those questions lead to different kinds of historical evidence. This path introduces three documents alongside an earlier astronomical mechanism. It helps you read the authors' arguments and proposed systems in their own settings before comparing them with later technology.

The dates below identify the historical publications or proposal. A digital transcription or PDF reprint can have a much later date and different pagination. Record the original title, author, date and section, as well as the edition you actually consulted. The documents are selected starting points, not a complete timeline of computing or a claim that one person invented every part of AI.

Questions to investigate

Step 1

Start with a device and the limits of reconstruction

Use the Antikythera essay to ask how surviving fragments support an explanation of a mechanism. Its linked research distinguishes observations of the remains from reconstructions of missing parts. Keep astronomical prediction, a general computing design and a discussion of intelligent behavior as separate tasks when comparing the examples in this path.

Choose one claimed function and trace it to the cited research. Record what survives, what the proposed reconstruction supplies, and which alternatives remain open. Avoid turning a present-day label such as computer into proof that an ancient maker was attempting every function associated with the word today.

Step 2

Lovelace's notes: distinguish a design from its uses

The 1843 English Sketch of the Analytical Engine presents Menabrea's account with notes by Ada Lovelace. Note A contrasts the Difference Engine's particular tabulation task with the wider proposed Analytical Engine. Note G develops a sequence for calculating Bernoulli numbers. The text is evidence of how the design and its possibilities were explained, rather than a record that every described calculation was performed on a completed machine.

Read the memoir and the translator's notes as distinct contributions. Choose one operation in Note G and identify its inputs, intermediate values and intended result. Then compare that specific explanation with a broad first-inventor label: explain which claim the document supports and what additional evidence the broader label would require.

Step 3

Turing's 1950 paper: examine the question and the argument

Computing Machinery and Intelligence appeared in Mind in 1950, volume 59, pages 433–460. Turing replaces a general question about machine thinking with an imitation-game question. The paper discusses digital computers, objections to machine intelligence and, in section 7, learning machines. Its arguments and expectations should be distinguished from measurements of a particular later system.

Read sections 1 and 2 before reducing the paper to a slogan about passing a test. Write down the participants, communication conditions and question the game is meant to investigate. Compare the discussion of Lovelace in section 6 with her own Note G. Label your comparison as an interpretation of two texts with different purposes.

Step 4

The Dartmouth proposal: read a plan as a plan

The proposal dated 31 August 1955 names John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon. It proposes a Dartmouth summer project for 1956, outlining topics including language, neuron networks, abstraction and self-improvement. Its central premise is presented as a conjecture. The proposal records an intended research agenda; it does not prove that the planned summer achieved every ambition.

Select one topic and separate its conjecture, proposed method and hoped-for result. To investigate what actually followed, seek dated reports, correspondence or later publications in addition to the proposal. An institution's retrospective history can help locate the event, while the original plan lets you examine what its authors expected beforehand.

Records and references to explore

These external resources support further investigation. Check each collection's scope, access conditions and item-level references.

Try a small source comparison

Make a four-column evidence table: the document's question, the mechanism or method described, what is demonstrated or merely proposed, and evidence needed for a later claim. Add a precise section reference to each row. Compare one idea across two documents without assuming an uninterrupted path from an old design to a current AI product.

Explore more history subjects →