Ewe, Welsh? Baa humbug

David Crystal describes his latest book as a "linguistic travelogue". Never having encountered this genre before, I did not know quite what to expect. Would it be tragedy, comedy, history, pastoral, pastoral-comical, historical-pastoral, tragical-historical or tragical-comical-historical-pastoral? The last, I guessed, but I was wrong.

Polonius's typology cannot accommodate a Crystal. By Hook or by Crook is autobiographical-whimsical-quizzical-oddsandendsical.

Because professional linguists so rarely let their hair down in print, especially when they have acquired the public status of an "authority" on the English language, this is a rare treat. The travelling concerns a field-trip Crystal undertook for the BBC's project Voices, the aim of which was to document the current state of spoken English. But there are also flashbacks to foreign climes where the seemingly indefatigable author has been in search of English. These include most of the countries of Europe, India, Australia, the US and even Zimbabwe.

The "linguistics" in the book is not too forbidding. There are some sparkling music-hall touches, none better than the opening scene in which Crystal looking to record Anglo-Welsh accents at a rural sheep market in Anglesey selects a likely-looking farmer from the throng, only to discover that his "informant" speaks broad Scots. I have heard other linguists recount similar tales of embarrassment, including one who got not just the wrong accent but the wrong language. It must be one of the occupational hazards of the profession. The Scots farmer at the Welsh sheep auction solemnly informs Crystal that he should really be interviewing the sheep, because they bleat differently from sheep in Scotland. Which leaves our hero not knowing whether he is having his leg pulled or not.

Do sheep have local accents? The question leads on to musing about Karl von Frisch's research on the so-called "dialects" of bees. This is typical of the way in which whatever happens along the way can tempt Crystal off into discussions of almost anything language-related that comes into his head.

Not knowing where it will lead next is one of the charms of the book, if you like that kind of charm. For those who need to feel that light reading improves their general knowledge, this is the ideal deckchair book: a potpourri of etymologies, place names, pub names, curious idioms and dialect expressions, all flavoured by personal reminiscences.

Crystal's subtitle prompts the question - not to be taken too seriously, of course - "Well, did you find it?" To stand a chance of finding something, you have to know what you are looking for, and then look in the right kind of place. Crystal looks mostly in dictionaries and, outside dictionaries, for speakers who speak any recognisable brand of English. But what exactly that is he never tells us, except that it's changing all the time. Perhaps it is so elusive that it is beyond even the most expert of linguists actually to define it: you just have to be able to "recognise" it when you meet it. That, at least, is one plausible interpretation of this book's message.

As with most travel writing, we eventually learn most not about the places visited but about the personality of the traveller. Crystal comes across as a mild, affable, erudite and patient man, a theatre-lover with a pleasant sense of irony and a weakness for watching downmarket television programmes. We are shown the traveller in various personae: Crystal the mad motorist, leaping perilously out of his Rover 75 to photograph a misspelt public notice; Crystal the academic, giving us a mini-lecture on the rights and wrongs of the great apostrophe muddle; Crystal travelling round the celeb circuit, delivering a talk here and a talk there, unfurling a flag on Shakespeare's birthday; Crystal the ludic linguist with his groan-making puns, who has to decide, when staying overnight in Stratford, whether to B&B or not to B&B.

Words, he seems to be telling us, are endlessly fascinating, good clean fun, capsules of interesting associations, but nothing to get worked up about. Language has its seamier side, but that is best avoided. Nothing too controversial (political or sociological) should be allowed to spoil one's enjoyment of it. On one occasion when a common "four-letter word" crops up in the discussion, Crystal refuses to spell it out in full, because - he tells us - he is not "in the business of upsetting readers".

Politically correct, as current academic fashion demands, a frosty Crystal duly disapproves of the linguistic stereotypes exploited by popular entertainers for comic effect. But one cannot in the end help wondering whether dialectologists, who go around with ears cocked to detect the least variation of vowel or consonant, are not themselves operating with, and thus perpetuating, the very notion of linguistic typecasting.

Roy Harris is emeritus professor of general linguistics, Oxford University.