Lost in translation: The iPhone's accents problem
"Whit did ye get her fir Christmas?"
"Ye firgoat? Aw, did she gie ye hell?"
"Well ye said ye firgoat"
"Naw, ah fir goat."
And so the pantomime joke continues, for as long as the colourfully clad dames can draw it out. Eventually the smaller, fatter ugly sister will understand that her taller, scrawnier stage sibling has given a present of a fur coat. But the confusion inevitably won't end there. "Whit fir?" The answer: "Fir tae keep her warm." Obviously.
Oh, they're funny things, vowels. Well, funny until we have to deal with speech recognition technology – as complaints about a new iPhone application's inability to recognise British accents have highlighted.
There's nothing more likely to wipe the smile off your face than sitting in public, trying to communicate with a machine unprepared to meet you even halfway.
"So you want to know how much money you have in your bank? Well, not until you've convinced everyone within earshot that you're a blathering lunatic who believes she can speak Whale language. Now say: one."
"I'm sorry, I did not understand you. Please try again."
"No. I'm still not getting you there. Try harder, human."
"Wooooon, wan, wooooooon, wuuun, won, woooooo. Oh fik/foook/fok this – where's the hash key?"
The Google application for iPhone, which was developed in the US, is supposed to allow users to search for information by recognising the words they say. Unfortunately, there have been some serious transatlantic translation glitches.
When someone with a Scots accent asked to search for "iPhone", the phone searched for "sex" instead, and provided a link to an adult website. On the second attempt, the search engine looked for "sledding".
Asked in a Kentish accent to search for "iPhone", it interpreted this as "my sister", then as "Einstein". One can only imagine the potential for chaos in Wales, where one person's accent led the machine to mistake "iPhone" for "gorillas" and "kitchen sink".
There's already enough potential for human misunderstanding when faced with different accents. Earlier this year, Scottish advocates were outraged after confused English stenographers were found to have transcribed Barlinnie as "Barrel Annie" and "libelled" and "fanciful" as "liable" and "fanciable", in errors put down to accent. So can technology ever hope to make sense of our varied pronunciation?
According to Jim Scobbie, professor of speech science at Queen Margaret University in Edinburgh, things should improve as computer technology expands to memorise more accents, although there is still, clearly, a long way to go.
But surely not even the most desperate human would mishear "sex" (or indeed sledding) when the word said was iPhone. Actually there could be a fairly simple explanation, suggests Scobbie: "The S consonant involves high-frequency sounds. If the phone wasn't carrying that frequency or there was a lot of background noise, the S could be mixed up with the F at the beginning of "phone", which is also a high-frequency sound."
Scobbie points out that the "I" sound has not been picked up in either case, but this problem, it seems, is not just a British/American one. When I phoned the university to speak to him, I was greeted by a voice recognition message that asked me to say the name of the person I wished to speak to. I was put through to a Jim Bain – 'nuff said. Or rather not enough.
The iPhone's recognition of "e" rather than "o" in the word phone is more difficult to explain, but may be down to Scottish accents' short vowels, which are particularly noticeable in that word, Scobbie says.
"English has a lot of vowels – up to 20, depending on your accent, and Scotland has only about 12, so that might be why Scottish people don't make the sound the phone expects," suggests Scobbie. The "e" sound in sex and sledding is a short vowel.
It does make you feel a bit marginalised when your accent isn't recognised, and you end up putting on a funny voice, says Scobbie, but it is just an unfortunate side effect of being a small country.
But the nation's gadget lovers shouldn't be too despondent. Scotland is considered a world leader when it comes to voice recognition, thanks to the Centre for Speech Technology Research at Edinburgh University. The prospect of millions of frustrated Americans doing their best Jean Brodie impression into their new mobile phones would make anyone smile. Altogether now, say "iPhone", everybody.