Interactive display
Interactive computing was in its infancy in 1966, and users interacted with ELIZA via a teletype.
Source: Wikimedia Commons

ELIZA is an early natural language processing computer program created from 1964 to 1966,[1] by Joseph Weizenbaum at the MIT Artificial Intelligence Laboratory.[2]

The program, consisting of about 200 lines of code,[3] was built in two layers, the first a language analyser. The second layer incorporated a script, which Weizenbaum likened to the rules “rather like those that might be given to an actor who is to use them to improvize around a certain theme”. In that way ELIZA could be supplied with different scripts allowing it to maintain conversations about “cooking eggs or about managing a bank checking account”. The name ELIZA was chosen because just like the Eliza Doolittle of Pygmalion fame, the system could be “taught” to speak increasingly well, by being supplied with updated scripts.[4]

ELIZA allocated a “precedence number” or ranking to keywords in the input sentence such as I, you, boyfriend, and then applied appropriate transformations such as turning “you” into “I“. The first and best-known script, known as DOCTOR, was designed to mimic the responses of a Rogerian psychotherapist;[a]The psychotherapist Carl Rogers was well-known for mirroring whatever the patient said, producing “an unusually constrained form of dialogue.”[5] Weizenbaum gave as an example the input “I need some help …” returning “What would it mean to you …”[3]

The original version of ELIZA was written in SLIP,[b]Symmetric LIst Processor.[5] a programming language invented by Joseph Weizenbaum.[5]

Response

Weizenbaum was disturbed by what he perceived to be the misinterpretation of his work on ELIZA, and in response wrote his book Computer Power and Human Reason: From Judgment to Calculation. He was particularly concerned that some practising psychiatrists believed that ELIZA could be developed into an almost mechanical form of therapy, and was “startled” to see how quickly some people interacting with the computer became deeply involved with it, the so-called ELIZA effect. In a more technical sense, he was also concerned that some people believed that ELIZA was a general solution to the problem of understanding natural language, which he believed was a problem that had no general solution.[6]

Notes[+]

References

  1.  
  2. pp. 3–4
  3.  
  4. p. 3
  5.  
  6. pp. 2–7

Bibliography

Bassett, Caroline. “The Computational Therapeutic: Exploring Weizenbaum’s ELIZA as a History of the Present.” AI & Society, vol. 34, no. 4, 2019, pp. 803–12, https://doi.org/10.1007/s00146-018-0825-9.
Ireland, Corydon. “Alan Turing at 100.” The Harvard Gazette, Sept. 2012, https://news.harvard.edu/gazette/story/2012/09/alan-turing-at-100/.
Shah, Huma, et al. “Can Machines Talk> Comparison of Eliza with Modern Dialogue Systems.” Computers in Human Behaviour, vol. 58, May 2016, pp. 278–95, https://doi.org/10.1016/j.chb.2016.01.004.
Weizenbaum, Joseph. Computer Power and Human Reason: From Judgement to Calculation. W. H. Freeman & Co, 1976.

External links