One of the heavyweights in this category is A Dictionary of Modern English Usage by H. W. Fowler. It was such a big hit when it first appeared that, according to the introduction to the latest printing, "within a few years, people no longer felt it necessary even to mention the title and talked simply of 'Fowler'" (vii). The first edition is still revered by language mavens.
A Dictionary of Modern English Usage was first published in 1926, almost 20 years after the idea for it was first brought to brothers Frank and Henry Fowler's attention. The dictionary was written during a rise in prescriptive grammar: language experts were telling speakers how they should speak, prescribing how language should work, as opposed to describing how language speakers used language. And their ranks were swelling. So a book that told you how to use this word or that phrase was just the thing. Henry and his brother Frank worked on their dictionary while James Murray was working on the Oxford English Dictionary, and they kept tabs on that project (check out The Meaning of Everything, which gives a history of the OED). Ironically, the OED's goal was, and remains, to describe all the words in the English language. It gives not just the current definition of a word but all previous definitions of that word back to its first appearance in print.
Frank and Henry started planning the book in 1911, though Frank was actively working on the Pocket Oxford Dictionary until his death in 1918 while fighting in the war. Henry then finished the POD and picked up his usage dictionary work again, finally completing it in 1926.
The dictionary was revised in a second edition in 1965 by Sir Ernest Gowers. I've never seen this edition, but according to the New York Times, Gowers "gave it a light going-over, preserving both the spirit and the substance of the original." The book was revised again in 1996 by Robert Burchfield. This edition is generally regarded as heavily edited and more of Burchfield's work than Fowler's. It was this edition I was introduced to and it worked well enough for me, until I found Garner's.
Now Oxford has seen fit to re-release the original Fowler, restoring it to its former glory, as it were.
So is it worth slapping down $20 or $30 for another book that will gather dust on your shelf?
First, Fowler shouldn't be the only usage book on your shelf, particularly if you’re writing for a US audience. Fowler is undoubtedly a man of his place. His dictionary covers modern English usage, not American usage. He’s also a man of his time. There’s a revealing, if outdated, essay on "feminine designations”"(175-176):
This article is intended as a counter-protest. The authoress, poetess, & paintress, & sometimes the patroness & the inspectress, take exception to the indication of sex in these designations … These ladies neither are nor pretend to be making their objection in the interests of language or of people in general; they object in their own interests only; this they are entitled to do, but still it is lower ground, & general convenience & the needs of the King's English … must be reckoned of more importance … With the coming extension of women's vocations, feminines for vocation-words are a special need of the future.Yikes. Did Fowler really say that we need words like murderess and edtriss (both listed in the entry)? Yup, he did. (Shudder.)
Fowler also reflects the his times in that he is sometimes prescriptive and sometimes descriptive. In his entry on split infinitives (558-561), he addresses people who don't know what a split infinitive is but who care about not splitting them:
These people betray by their practice that their aversion to the split infinitive springs not from instinctive good taste, but from tame acceptance of the misinterpreted opinion of others; for they will subject their sentences to the queerest distortions, all to escape imaginary split infinitives ... the havoc that is played with much well-intentioned writing by failure to grasp that distinction is incredible ... After this inconclusive discussion, in which, however, the author's opinion has perhaps been allowed to appear with indecent plainness, readers may like to settle for themselves whether, in the following sentence, "either to secure" followed by "to resign", or "to either secure" followed by "resign", should have been preferredFowler concludes that split infinitives are grammatical in English and should be allowed.
But at "elemental, elementary" (133), Fowler describes the situation without making comment:
The two words are now pretty clearly differentiated, the reference of -al being to "the elements" either in the old sense of earth, water, air, & fire, or as representing the great forces of nature conceived as their manifestations ... & that of -ary being to elements in the more general sense of simplest component parts or rudiments.Fowler does, indeed, have wisdom to impart, a wisdom that has held up for almost 85 years. In addition to the split infinitive conclusion (one largely held now to be the correct position), check out the entry for "affect, effect" (13):
These verbs are not synonyms requiring differentiation, but words of totally different meaning, neither of which can ever be substituted for the other. Affect ... means have an influence on, produce an effect on, concern effect a change in: effect means bring about, cause, produce, result in have as a result.It's worth noting here that Burchfield kept this entry in his version but substituted his own essay for the split infinitives entry -- and came to a different conclusion.
No one usage manual can do all things. Have more than one. Compare one's advice to another's and think about what truly lies in these entries. Fowler's classic first edition should be one of those -- for its advice that stays current and for the place in language history it illustrates.