Hi, As it happens my PhD topic in 1967-70 was speech recognition (not natural language processing, but the topics overlapped). I just checked the list of references in my thesis for the first time in rather a lot of years, and the only relevant citation seems to be: Lindgren, N., Machine recognition of human language, IEEE Spectrum, March 1965, p. 114 and April 1965, p. 45. I don't have copies, but I assume that a Spectrum article would include references to current work. Note that the McConlogue/Simmons paper refers to quite a bit of previous work, so I suspect that as for most topics, Zeitgeist is a big part of the story. For example, the paper says "The system was tested on 400 elementary basic English sentences including 300 used earlier by Knowlton in a different learning parser." So no, this paper wasn't first-in-field. You'll also note that its dataset was trivial compared to what a modern LLM would learn from. In the mid-1960s, approaches to language understanding were very much grammar-based, which is of course absolutely not how LLMs work today, and the McConlogue/Simmons paper is typical. I'm pretty sure I didn't see it at the time, or it would have been cited as background in my thesis. There was also quite a lot of work on Chomsky-style transformational grammar, for example in the Computational Linguistics Project under Joyce Friedman at Stanford CS. But ultimately grammar-based approaches were not the answer. I think it's fair to say that the McConlogue/Simmons paper was leading-edge at the time. The authors were at SDC, which suggests Defense money (or should I now say War money? Will DARPA now be called WARPA?). Regards/Ngā mihi Brian Carpenter On 12-Sep-25 07:18, Hansen Hsu via Members wrote:
Hi all,
A woman in San Francisco recently reached out to CHM asking if a paper her mother, Keren McConlogue, co-published in CACM in 1965 is the first use of machine learning in a natural language parser, and what place it has in the history of machine learning and NLP.
/Analyzing English Syntax with a Pattern-Learning Parser/ Analyzing English syntax with a pattern-learning parser | Communications of the ACM <https://dl.acm.org/doi/10.1145/365660.365683> dl.acm.org <https://dl.acm.org/doi/10.1145/365660.365683> 365660.cover.jpg <https://dl.acm.org/doi/10.1145/365660.365683>
<https://dl.acm.org/doi/10.1145/365660.365683>
I can’t find anything earlier than this. Has anyone looked into the earliest uses of what would have been called “pattern recognition” in building a parser that could construct a syntactic grammar by learning from example sentences?
.........................................................
*Hansen Hsu, Ph.D.* *Curator, Software History Center* *Computer History Museum *he/him**
1401 N. Shoreline Blvd. Mountain View, CA 94043 (650) 810-1880 direct (650) 810-1010 main (650) 284-6254 mobile (650) 810-1055 fax hhsu@computerhistory.org <mailto:hhsu@computerhistory.org>
Follow Us: CHM.org <https://computerhistory.org/>__IFacebook <http://www.facebook.com/computerhistory>__ITwitter <http://www.twitter.com/computerhistory>__IYouTube <http://www.youtube.com/computerhistory>__ISubscribe to Our Newsletter <https://www.computerhistory.org/subscribe>
_______________________________________________ This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. Opinions expressed here are those of the member posting and are not reviewed, edited, or endorsed by SIGCIS. The list archives are at http://lists.sigcis.org/pipermail/members-sigcis.org/ and you can change your subscription options at http://lists.sigcis.org/listinfo.cgi/members-sigcis.org