Skip to 0 minutes and 14 seconds Well, it’s a great pleasure today to be able to talk to Michael Rundell, editor-in-chief of Macmillan Dictionary, macmillandictionary.com, whom I’ve known for many years. And you’ve spent a life, I think, looking at words and lexicography. So when did you begin? Well, almost prehistoric times in dictionary terms. I got into lexicography in 1980, which was just at the dawn of the corpus era. Or, really, in my case it was at the end of the pre-corpus era, because I did spend a couple of years at the beginning of my career in dictionaries working without corpora.
Skip to 0 minutes and 54 seconds So the way that people used to– you’d have citation forms, you would look at what other dictionaries were doing, and you would tap into your own mental lexicon, intuitions, and so on. So that was the way things were done in 1980, but of course, things changed very quickly after that. After I’d done a couple of years with one publisher, I moved to Birmingham University to the COBUILD project. Well, that must have been very exciting. Which– it was, it was. It was ‘82.
Skip to 1 minute and 27 seconds It was a very exciting time. This was when the first corpus based dictionary was being developed. And thought through, I suppose? You had no model. It did have no model, and that was the interesting thing about COBUILD. It was very exciting. You were on a daily basis making new discoveries from the corpus about how words and phrases worked and so on. But on the other hand, there was a sort of element of shambles about it because, inevitably, it was a whole new discipline really. Corpus lexicography was in the process of being born.
Skip to 2 minutes and 8 seconds I worked there not for very long, just for a couple of years, and then moved to Longman where I was in charge of dictionaries for quite a long time. We’ll talk about Longman in a moment, but when you started using corpora to create dictionary entries, was there a particular eureka moment? Where you thought, this is the way it should be done. For myself, there was a eureka moment when I thought, this is definitely the way to do the study of language or a really important way. Was there a moment for you like that? Yes.
Skip to 2 minutes and 39 seconds You would be given– you would be assigned a set of words to basically create dictionary entries for, and a bunch of corpus data, which in those days was in the form of concordances on paper. Right, printed out. We didn’t have– On a dot matrix printer or something. Probably, yeah. Yeah, you would start looking at this stuff and it would be kind of, wow, almost several times a day. I do remember a word– talking of eureka moments– I remember looking at the word “represent” and spotting something which didn’t seem to have been covered in the dictionary record up to that point.
Skip to 3 minutes and 26 seconds If you sit down and think about “represent” if you introspect, you sort of think about one thing represents something else. Or people represent a constituency in parliament, that sort of thing. Or this sign represents– this road sign represents something or other. But what we spotted was, what was coming through very clearly in the concordances, was this example where people would say in newspapers things like, this represents an enormous breakthrough in the conflict between x and y. Where “represent” is really, just a copula verb it really means “is” fundamentally. And it didn’t seem to have been captured in any of the record. And yet, it was very striking, because there’s plenty of it in the corpus.
Skip to 4 minutes and 18 seconds And then you just think, wow, there’s a whole– not a new meaning– a meaning which people have been, a form of the word people have been using– Perfectly happily. –but that just was– sort of evaded anyone’s notice, because it was just– Below the level of consciousness. In a way, yeah. Of course there’s so much of that. So yeah, there were lots of moments like that working in COBUILD.
In Conversation with Michael Rundell
Professor Tony McEnery speaks with Michael Rundell about Corpus Linguistics and dictionary construction
© Lancaster University