One language among many

I first used speech-recognition software in 2012. I was about to begin work on my novel The Crying Place, and I wanted to get closer to oral forms of storytelling – a sitting-around-a-campfire feel to the text. Also, I was intending to draft much of it on the road from Naarm (Melbourne) to Mparntwe (Alice Springs), then later in the desert, a moving vehicle not being the easiest place to type.
The software took a bit of getting used to. I had to train it to decipher the idiosyncrasies of my voice by reading texts out loud and using it in different environments. And I had to get the hang of directly narrating the story in character – a form of oral ‘method writing’… What soon became clear was that sounds other than my voice, such as birdsong and wind passing through trees, were being picked up by the software and ‘translated’ into words. Mulga came up as ‘will will’, desert oak as ‘with with’, birds varying according to song. At first I deleted these words from the text – treated them as errors or intrusions – but gradually I recognised what they were showing me. The direction the wind was travelling. The proximity of birds and their conversations, sometimes interspecies. Other languages.