On the Perils of "Best Book" Lists

The Guardian has completed a list of "the hundred best non-fiction books of all time" which illustrates, by its sheer badness, the perils of doing any such thing.

To begin with, they also interchangeably use the term "greatest" in place of "best", and the two are not the same. "Great" remains tied to its root meaning of "large" even in this sort of context, and while we expect that great books will be good, we have lots of conceptual categories allowing for good / better / best in ways that are not great. Great is a narrower field.

Secondly, the selection is heavily conditioned by the issue of influence. Books have been chosen purely because they have been influential. This category has members which overlap with "best" and "greatest" in fact but not conceptually. For example: Dr. Spock's book of baby and child care is on the list. It was certainly influential, but is in no way an obvious candidate for "one of the best non-fiction books" and indeed has probably already been displaced for most parents by other manuals ("What to Expect..."). It is not likely to continue to be read for either its deathless prose or for the continuing and unshaken reliability of its contents. Likewise, Dale Carnegie's How to Win Friends and Influence People may be influential but it is a massive stretch to consider it the best of anything, even the category of self-help books.

Thirdly, the list restricts itself to works originally written in English, and Modern English at that: nothing older than the Seventeenth Century. Asser, Aelfric, Chaucer, and Elyot need not apply. (In fact, it lets in one work violating this restriction: En Attendant Godot was written before Waiting for Godot.) This blows a large hole right below the waterline of the whole enterprise: if you're going to put Hume's Treatise of Human Nature on the list, there is no way other than an arbitrary cut-off to exclude the Nicomachaean Ethics or A Critique of Pure Reason. At one fell swoop you exclude every one of the Anciens; Dante and Aquinas are put on the trash-heap, as are the great works of the German and French Enlightenment. And the title does say "of all time" and does not say "written in English".

In comments to follow, these arbitrary restrictions of time and language will be ignored, taking "best non-fiction books of all time" at its face value.

Finally (before getting down to details), instead of doing the intelligent thing and reserving judgement on books published in (say) the last 50 years on the basis that proximity impairs our perspective, the list is heavily weighted in favour of very recent books. 15 come from the last 50 years and fully 50 of the 100 from the last hundred. (And it's especially among those most recent books that many of the more dubious works pop up: will Dreams from My Father be considered a classic autobiographical work in 100 years time?) Quite aside from questions of perspective, it stretches credulity that so many of the best/greatest books are so recent.

The list also stumbles in determining its ambit by subject (in my opinion) by including poetry - a subcategory with the specific drawback that it's one if the oldest categories of written work, making the exclusion of the Ancients and Mediaevals far more serious: not only do Herodotus, Aristotle, Plato, Thucidides, Josephus, Augustine and Aquinas have to be air-brushed out, but so do Homer, Sappho, Aeschylus, Sophocles, Euripides, Virgil, Ovid, Horace, Petrarch, and Boccacio. (Dante and Chaucer wrote prose treatises, but their major works would be included only as poetry.) In addition, it allows fiction to sneak in the back door: the First Folio of Shakespeare has more in common with Fielding or Austen than it does with Robert Burton.

At a detailed level, the fact that so many different categories are represented makes judgement frequently like comparing chalk and cheese. How to talk about the relative importance of a cookbook versus a book of literary criticism? Within a category, though, the inclusion of even one work immediately raises the question of other works within that category.

Thus, the only work of literary criticism on the list is F. R. Leavis' The Great Tradition. But it is not unreasonable to consider at least some of Aristotle's Poetics, Coleridge's Biographia Literaria, Bradley's Shakespearean Tragedy, Auerbach's Mimesis, Curtius' European Literature in the Middle Ages, Empson's Seven Types of Ambiguity, and Frye's Anatomy of Criticism as being better candidates - better written, better argued, more informative, and even arguably more influential. (And that's not even looking at classical studies, excluded from the list except for Gibbon, with its own set of great works.)

The list is relatively heavy in autobiographies but excludes the Confessions, the very font of the form, and Montaigne's Essais, two books whose absence from any "top non-fiction" list is inexplicable.

The list includes some strong candidates in history, along with some weak ones: to my taste there are some omissions which are arguable either way - I would put in Wedgwood's history of the Civil War and Huizinga's The Autumn of the Middle Ages (and if I were to invade the last fifty years, Hobsbawm's The Invention of Tradition, Duffy's The Stripping of the Altars, and Eisenstein's The Printing Press as an Agent of Change). There are also some which simply should not be excluded, period, such as Thucydides The Peloponnesian War and Caesar's Commentaries.

As for science: is Hawking's A Brief History of Time more deserving of a place than Newton's Principia (or, for that matter, Russell and Whitehead's)? Schrödinger, Feynmann, D'Arcy Thompson, Lyall, and Einstein also all should have a look in.

Philosophy is ridiculously weak: Hobbes, Locke and Hume are there, but none of Plato, Aristotle, Augustine, Aquinas, Descartes, Kant, Hegel, or Wittgenstein are present.

Some entire fields are skipped over. I've already noted classical studies, hanging in by a fingernail because of Gibbon; likewise, linguistics is barely represented by Mencken's The American Language. But sinology, geology, theology, textual criticism, and many others are simply not represented at all.

It's possible to argue that there is no objective ranking possible allowing one to speak at all precisely of "the best books" even in a given class. (Some people would deny any such classification is possible at all, except by taking some specific criterion rather than a general better / worse, but I'm reasonably happy to say that The Making of the English Working Class is simply better, as a book, than Herman's The Cave and the Light, so I'm not going to rely on that line of argument.) But even ranking within a discipline inevitably has many blurry lines, and comparing between disciplines is essentially comparing incommensurables.

If you avoid putting a numeric limit like 100 on inclusions it might be just possible for a polymath, with hubris, to think of assembling a list of "the best books of all time", drawing borders a little wide for safety (though my selection would still drop some of the contents of the Guardian list). Doing it with a limit is simply folly. This list doesn't even get close.