Some time in the fall of 1985, I contacted Alan Prince, who had occupied the adjacent office when I was postdoctoral fellow at the Center for Cognitive Science at MIT in 1979-80. I needed a partner for a debate at the Center against James McClelland, who had recently presented a revolutionary neural network model of the acquisition of the past tense in English. McClelland and his collaborator David Rumelhart had claimed that their model made generative linguistics, of the kind practiced by Alan, obsolete, and I knew that Alan would be well-equipped to spot the shortcomings of the model and defend the power of the generative approach.
After a lunch on a cold and rainy November day, Alan sent me a note via the new medium of email that would redirect my life. In that memo he went well beyond a critique of the model. I could see that in those few notes he laid out a few simple but profound phenomena that could establish the scientific rationale for phonology, phonetics, morphology, mental representations, modularity, symbolic rules, universal grammar, and the distinction between memory and computation. Within minutes I realized that these phenomena, studied as topics in experimental psychology, would set the agenda for the next phase of my research career.
Thus began a collaboration that went well beyond the debate at MIT and our critical analysis of the Rumelhart-McClelland model published in Cognition in 1988 (which went on to garner 500 citations, my second-most cited article). Between 1985 and 1992, when Alan left the Boston area for Rutgers, we coauthored an additional five articles, co-taught a course, and spent sabbaticals at each other’s institutions. And Alan’s influence went well beyond our direct collaboration. The ideas in his original memo would serve as the inspiration for four grants, a book (Words and Rules: The Ingredients of Language, 1999), and seventeen substantive journal articles, the most recent published last year, almost thirty years after our initial meeting. The topics could not be more diverse: children’s speech errors, cross-linguistic comparisons, historical changes in English, Big Data corpus analyses (including the article in Science that introduced the world to the Google ngram viewer), studies of the neurobiology of language using fMRI and implanted electrodes, and the breakdown of language in people with brain lesions, Alzheimer’s, Parkinson’s, and Huntington’s Diseases, and genetically caused Specific Language Impairment.
Over the course of several moves and computer crashes I have lost Alan’s original memo, but most of the content was reproduced in the slides for a talk we gave in October, 1988, presented here in the original pre-Powerpoint source file. Also unearthed here is the oldest email I could find from me to Alan (dated January 20, 1987, and sent to the pre-Internet address cogs@alan%brandeis.csnet%csnet-relay), while we were exchanging drafts of the Cognition paper, together with the oldest email from Alan to me that I could find, dated October 5, 1989. They capture the intensity of the most important collaboration of my career, and embody the epigraph to our first paper which Alan had excerpted from Robert Frost: “If design govern in a thing so small.”
Pinker, Steven. 2015. A Consequential Collaboration. In Short ’schrift for Alan Prince, compiled by Eric Baković. https://princeshortschrift.wordpress.com/archives/pinker/.