I entered graduate school in the field of Experimental Psychology and began looking for a ‘killer topic’ for the master’s thesis. I knew I couldn’t just write a paper – it had to be a research project. This meant coming up with a theory and collecting data to support it. Furthermore, it had to be interesting enough to get published in a professional journal. I had no idea what I was going to do but I didn’t have much time. While preparing for a presentation I was giving at a seminar in learning theory, I ran across a study showing how reading scores dramatically improve once children start making ‘inferences’. For instance, when a child reads the sentence “The paratrooper jumped out of the door” it helps if they make the inference “The paratrooper jumped out of the plane”. We weren’t teaching children to make inferences and that was one of the reasons why Japanese students were outscoring American student in reading. That interested me so I made a copy of it for future reference. Later that year at a seminar in linguistics, I listened to one of the presenters describe how inference-making can have undesirable consequences. At the Watergate hearings for instance, when President Nixon said “No one told me John Mitchell was behind the break-in.” People came away with the impression he said, “I did not know John Mitchell was behind the break-in.” I thought, how ironic ..the same process that helps children learn to read can become a source of misunderstanding between adults. It was an intriguing theory but even more important – I had found a ‘killer topic’. After getting my thesis committee to approve, I began running tests measuring adult speech and reading comprehension. What I found is that people really are susceptible to misleading inferences. Over two-thirds of the time, they mistake the inferences they make for something they either heard or read before. After listening to a passage containing the sentence “The karate expert hit the cement block”, they swore what they heard was “The karate expert broke the cement block.” Relying on inferences can actually lower adult reading scores. I thought I was on to something. Inferences are stored in a manner that is indistinguishable from information conveyed by direct assertion. I found examples occurring in courtroom testimony and deceptive advertising. I presented my findings at a conference of the Western Psychological Association in Los Angeles where I managed to get my thesis published [link]. Then, someone approached me and said that my research would be useful in another field .. developing speech recognition systems for the computer. So in 1981, after the Reagan Administration began cutting funds for the kind of research I was doing, I completed the master’s degree and I went to work for the computer industry. I remember my colleagues telling me I was ‘selling-out’. I never lost interest in psychology and always thought I’d come back and complete a doctoral degree.
