Tom Lubbock: Don't take this book as the last word on art. It can be only an introduction

Some people look to the arts, not for an expanding horizon, but for closure. So many works, so many choices, so many opportunities for getting it wrong ... How much simpler to have it all sorted, graded, and listed – to have the best, and in the bag.

So you get boxed sets called the like of The Only Opera Collection You'll Ever Need and The Greatest Classical Music Album Ever. Or publications called The Folio Society Book of the 100 Greatest Paintings. And, taken at face value, they seem a sorry business. But it depends, of course, how they are used.

It depends whether they're the end or the beginning of a story. If you really did get this guide to the 100 greatest Western paintings in the hope that you would never need to encounter any other painting, that would be your bad luck. But as someone who first dipped his adolescent ear into classical music through the CBS series of Greatest Hits cassettes (Bach's Greatest Hits, and so on), I know it can work out otherwise. The best way to assess this selection is not as alast judgement, but as a starter pack.

Besides, it obviously isn't an attempt to choose the 100 greatest paintings. If you were doing that, some artists would appear more than once. (It could well be that the 100 greatest paintings were painted by no more than 25 artists.) But the Folio Society has a rule of one picture per artist. Greatness gives way to other considerations, such as variety and chronological spread (Duccio through to Picasso).

In which case, it doesn't really matter how great the chosen pictures are. I might protest that neither Benjamin West's Death of General Wolfe nor John Singer Sargent's Carnation, Lily, Lily, Rose should find a place, even among the 1,000 greatest paintings. But then, one or two duds in the mix are all right. Presuming the reader doesn't want to take this selection as the be-all-and-end-all of art, the odd opportunity not to admire is welcome.

On the other hand, as the book is intended for beginners, some guidelines seem useful. For example, to fill it up with the most famous pictures would be pointless. People are going to bump into Munch's Scream, Klimt's Kiss, Botticelli's Birth of Venus, Michelangelo's Creation of Adam and Bruegel's Hunters in the Snow anyway, in the windows of poster shops. A little adventurousness, even eccentricity, is wise. It's good, for example, that Georges de la Tour is representednot by his famous, dreamy, candle-lit visions, but by one of his sharp-edged card-cheating scenes.

What's more, the choices should inspire further exploration. What's depressing about having Caravaggio represented by his early, static Lute Player is not only that it's far from his best, but that this choice conceals from the novice the terrific excitement of Caravaggio's art. It does not truly introduce.

Overall, this selection is a good as any. But, actually, for a really engaging introduction, you don't want a book of the 100 greatest paintings. Greatness glazes. You want a book of the 100 worst – or most overrated – paintings. That would get people looking.