Thinking Machine
A thinking machine is the long-standing idea of an artifact that does what minds do, reasoning, judging and perhaps creating, and the phrase carried a charge of wonder, fear and skepticism for a full century before the term artificial intelligence existed. This entry follows the phrase through its first serious technical discussion, in the 1843 notes on Charles Babbage's Analytical Engine, to Turing's 1950 reply and the popular press language of the 1940s and 1950s, ending with how the wording keeps reappearing in current technology marketing and debate.
The first careful discussion came from a translation, not a manifesto. In 1842 the Italian engineer Luigi Menabrea published in French a description of Babbage's proposed Analytical Engine, a programmable mechanical calculator driven by punched cards, and in 1843 Ada Lovelace translated it for Taylor's Scientific Memoirs and appended notes, labelled A to G, that were longer than the original text. Note G sets out in detail how the engine could compute Bernoulli numbers. Lovelace also suggested the engine's operations might act upon things other than numbers, for example the notes of music, if relations among them could be expressed in its notation. The engine itself was never completed, so every claim about it was a claim about a design.
Lovelace's caveat became the canonical objection. In the same notes she wrote that the Analytical Engine has no pretensions whatever to originate anything and can do whatever we know how to order it to perform. She was guarding against exaggerated claims about the machine's powers, not ruling out future mechanical intelligence, but "Lady Lovelace's objection" became shorthand for the thesis that a machine can never be creative or surprise its programmer. Douglas Hartree repeated the passage in his 1949 book Calculating Instruments and Machines, which is how Turing met it.
Turing's reply, in section six of the 1950 paper, challenges what the caveat actually shows. He pointed out that Lovelace and Babbage had no way to anticipate the storage and speed of later machines, and he reformulated the objection as the claim that machines can never take people by surprise. His answer was that machines surprise him constantly, since a person writing a program cannot track all its consequences, and that being surprised is a feature of the observer's mental act and not of the machine's inner nature. He also separated the idea that a machine must be told every step from the idea that it must be told every outcome, a distinction that anticipates programs that learn from examples.
In the 1940s and 1950s the popular press pushed the phrase into everyday speech. Newspapers and magazines described the new electronic calculators as giant brains, and the engineer and actuary Edmund Berkeley gave the idea a book-length treatment in 1949 with Giant Brains, or Machines That Think. The neurosurgeon Geoffrey Jefferson answered the enthusiasm with a skeptical Lister Oration in June 1949, and Norbert Wiener's 1948 Cybernetics linked machines and nervous systems in the scientific literature. Within this climate, "thinking machine" could mean an enthusiastic forecast, a headline convention, or a target for ridicule, and the ambiguity was part of the appeal.
The phrase echoes today, with the same two flavors. Marketing copy for AI systems speaks of models that think, reason or understand, and many systems now emit visible intermediate reasoning before an answer. Critics respond with Lovelace-like warnings: the system produces outputs from learned statistics and training data, and should not be credited with intent. The debate in both directions tends to treat "think" as one yes-or-no property, though Turing's practice was the opposite, to break the question into specific abilities that could be tested.
Misconceptions run in both directions. Lovelace is sometimes cited as proof that computers cannot be creative, but she was discussing one unbuilt engine and also imagined it composing elaborate music. She is also sometimes portrayed as the sole author of the program in Note G, whereas Babbage had worked on earlier programs for the engine, a point historians dispute in degree. And the press language of giant brains was not a considered theory of mind, only a vivid label that outran what the early machines could do.
Read "thinking machine" as a phrase that has meant something different in each decade it was used: a design caveat in 1843, a headline in 1949, a research question in 1950, and today a claim to be tested capability by capability.