English Language handbook

Contributed by:
This book covers syntax, punctuation, style, organization, and tone. A special feature of the book is a combined glossary/index, alphabetically arranged to give instant answers to the most commonly asked questions about misused words, phrases, and constructions, and cross-reference to the text if a longer explanation is required.
5. THE
7. THE
Edward D. Johnson
New York • Oxford
8. The Handbook of Good English: Revised and Updated
Copyright © 1983, 1991 by Edward D. Johnson
All rights reserved. No part of this book may be reproduced or utilized in
any form or by any means, electronic or mechanical, including photocopying,
recording, or by any information storage or retrieval systems, without permission
in writing from the publisher. For information contact:
Facts On File, Inc. Facts On File Limited
460 Park Avenue South Collins Street
New York NY 10016 Oxford OX4 1XJ
USA United Kingdom
First published as The Washington Square Press Handbook of
Good English by Pocket Books, a Simon & Schuster division of
Gulf & Western Corporation, 1230 Avenue of the Americas,
New York, N.Y. 10020
Library of Congress Catalog Card Number: 91-21872
A British CIP catalog record for this book is available from the British Library.
Facts On File books are available at special discounts when purchased in bulk quantities
for businesses, associations, institutions or sales promotions. Please call our Special Sales
Department in New York at 212-683-2244 (dial 800\322-8755 except in NY, AK or HI) or
in Oxford at 865/728399.
Manufactured by the Maple-Vail Book Manufacturing Group
Printed in the United States of America
10 9 8 7 6 5 4 3 2
This book is printed on acid-free paper.
9. <* CONTENTS k>
The Sentence 3
Case of Nouns and Pronouns: Subjective,
Objective, and Possessive 20
Agreement 32
Verb Tenses: Past, Present, and Future 48
Verb Moods: Indicative, Imperative, and
Verb Voices: Active and Passive
Modifiers 65
Sentence Structure 83
Comma 94
Semicolon 121
Colon 125
Dash 131
Parentheses and Brackets 136
Question Mark 143
Exclamation Point 149
Quotation Marks 151
Points of Ellipsis 166
Apostrophe 172
Hyphen 182
Diagonal 216
10. • Contents
Numbers 222
Dates 228
Abbreviations 229
Generic Terms 232
Titles of Officials and Names of Their Offices 234
Forms of Address 238
Place-names 241
Titles of Publications and Works of Literature,
Works of Art, Musical Compositions, and
Other Works 245
Foreign Words 256
Occasion and Intent 262
Organization 267
Tone 271
Revision 278
11. <^ PREFACE K>
This book's first edition was published in 1982. Only seven
years later I decided to revise it, but not because I thought it
had so quickly become out of date. It was based on more than
twenty years of experience as a book editor and more than half
a life spent largely in well-spoken company and I didn't think
either it or I was substantially dated. I was aware of some new
uses and misuses of the language and wanted to comment on
them, but my primary intent was to rectify shortcomings that
had been exposed by seven years of testing the book against
writing I had edited or read for pleasure and speech I had heard.
I wanted to expand my discussions of many details, modify my
judgments on a few matters, increase the number of cross-
references, and enlarge the Glossary/Index—all of which I have
In the course of the revision, however, I discovered that
English and attitudes toward it have changed more than I had
thought, and that I have changed too.
For one thing, the language has made adjustments to com-
plaints that it is sexist, and it continues to adjust. I discuss this
change and my accommodations to it under sexism in the
Glossary/Index; it has affected the diction in this revision
considerably. In 1982, I think, avoidance of sexist diction
would have weakened my book for many readers, but now,
sexist diction would weaken it, because genderless expressions
that once were evasive and obtrusive have become straightfor-
ward and unsurprising.
Another change—perhaps it is partly an effect of the swift
and broad acceptance of nonsexist alternatives to traditional
diction, which has demonstrated the adaptability of English—
is an increasing awareness among those interested in language
and correct use of it that correct is not always easy to define. In
the 1970s, several widely read writers on language came down
12. • Preface
heavily on usages and constructions that they considered de-
based, inane, despicable—and these writers' readers tended to
accept such condemnations humbly, even guiltily As the 1980s
began, so did an antithesis in popular writing on language. The
best-selling "prescriptivists" of a few years before were rebuked
for their bad temper and often jeered at for their bad schol-
arship. The "permissivists" insisted that English was what it
was and would change as it would.
Now we are perhaps in a lull in the war between prescrip-
tivists and permissivists—or in a battle of that war. The war
has been going on for centuries, and the current battle may
have been evident in the broad world of letters only in the past
decade but has been in progress in smaller arenas for some
time, certainly since the publication of Webster's Third New
International Dictionary in 1961 (an event discussed under
usage in the Glossary/Index). But if there is a lull, nevertheless
consciousness has been raised. The broad writing, speaking,
and reading public is now not so easily cowed.
The first edition of this book—though ''strict," which is to
say prescriptive—was considerably more genial in tone than
many similar books of its time, and, unusual for prescriptive
books, it did its best to explain its prescriptions or admitted
that there was no explanation but convention. However, it took
it for granted that any reader consulting it would share its
author's belief that there was such a thing as "good English"
and that it was worth learning.
The present edition is as strict as the first. It assumes that
those who use it want to be protected from criticism—and
there are still plenty of critics. The general culture may have
become more permissive about language, but that does not
mean there are no more critics; in fact, the polarizing effect of
the prescriptivist-permissivist battle has probably both in-
creased their number and hardened their opinions. And—in my
view—a great many of their opinions remain right, if there is
such a thing as good English.
This edition does, however, take even more pains than the
first to explain its rules and to distinguish logic from tradition,
tradition from prejudice, prejudice from common sense, com-
mon sense from nonsense. It is more thoughtful and, I hope,
wiser; it has been through the battle. And as its author, I feel
obliged, as I did not in 1982, to explain at some length what I
mean by good English, why I feel qualified to expound on its
strictures, and why I believe learning those strictures is worth-
13. Preface •
Good English changes over the course of time, and at any
given time there is some disagreement about what it is, both as
a concept and as an accumulation of usage details. I begin my
definition with a statement that may be self-evident but should
make it clear that the advice in this book, though "strict," is
not based on absolute truths: Good English is English that at
present very rarely sparks the expressed or unexpressed reac-
tion "That's not good English/' either from those who really do
know better or from those who merely think they do. I say
"very rarely" rather than "never" because usage arbiters don't
always agree, and also because critical reactions of two kinds
cannot be avoided. On the one hand, the reactions of those who
know almost nothing can be entirely wrongheaded and must
sometimes be ignored. For example, I have been criticized for
saying between her and me on the ground that between she
and 1 is more elegant—but elegant or not, and I say decidedly
not, between she and I is wrong. On the other hand, the
reactions of those who know almost everything, the true, and
few, serious scholars of language and usage, can be excessively
rightheaded. For example, careful avoidance of plural pronouns
such as their after singular pronouns such as everyone is justly
criticized by the truly knowledgeable as a rejection of a natural
usage that has been common in the best literature for cen-
turies. But a much larger minority, those who are not scholars
but do in general "know better," reject the usage, so I think we
must reject it too.
To continue my definition, good English is a kind of snob-
bery. It is not standard English but the English of a minority
who are likely to consider themselves superior, and are also
likely to be considered superior by others. English that is good
enough in one context may not be good enough in another, and
thus good English amounts to savoir faire, a touchstone of the
snob. All of us fail to use it occasionally, and some of us fail to
use it frequently. Those who fail infrequently look down on
those who fail frequently; those who fail frequently either live
in constant fear of embarrassing themselves or find some way
of taking pride in their unvarnished expression. Those who fail
infrequently make further distinctions among themselves; the
famous grammarian H. W Fowler observed, "Almost every man
is potentially a purist and a sloven at once to persons looking at
him from a lower and a higher position . . . than his own."
Grammar and usage are therefore touchy subjects, like class
distinctions—they are class distinctions. We expect occasional
correction from a parent or teacher, but any friend who cor-
14. • Preface
rects us had better be a good friend indeed; he or she is im-
plicitly criticizing our background, our education, our place in
the world, our being. And though many of the strictures of
good English promote clear expression and clear thought,
many others are merely the prejudices of language snobs. Con-
sequently, those of "good" background are frequently in a posi-
tion to criticize a speaker or writer who has not snared their
advantages but may have superior intelligence and superior
overall command of English. Such criticism is unfair and un-
democratic, but also far from uncommon; it is simply a fact of
society. In this book I usually identify strictures that are preju-
dices, and so readers who are not snobs and are immune to
snobbery can choose to ignore them—but I think few of us are
entirely unsnobbish or entirely immune to snobbery; I am not.
Longtime editors like me are, however, at least relatively free
of language snobbery. We spend our days and years correcting
the written expression of others, some of whom we are forced
to recognize as more intelligent, more highly educated, more
sophisticated both socially and verbally, and more successful
than we are, and unless we are unusually ill-natured we even-
tually are led to admit to ourselves that our skill is a humble
one and that those we correct often have much more to express
than we do and often express it with much more flair than we
could. We allow superior writers many liberties. It is likely that
every so often we have been slapped down by such writers for
making ill-considered changes, and we have learned from our
humiliations. We have a massive armament of arbitrary pre-
scriptions and niceties, but we bring the big guns to bear
chiefly on mediocre and bad writing—which improves mark-
edly when so attacked, partly because editorial routines often
expose faulty thought, which can then be attended to; our
skills do have an important function in this wordy world.
We find it difficult to explain our weathered, dispassionate,
and sometimes permissive attitude to friends who think we
should be "guardians" of the language, and who may use En-
glish carefully and well but resist its natural evolutions and
hold passionately to usage prejudices that they cannot justify.
We do very often impose such prejudices on what we edit, since
we want to protect those we edit from criticism both right-
headed and wrongheaded, but we may not share them. We
know the rules, we know the prejudices, but the responsibility
we have assumed as professional meddlers, accountable for
what we do, has made us respectful of the expression of others.
15. Preface •
We also, of course, have our private feelings about English
and its proprieties, just as do all users of the language. Our
professional experience entitles us, I think, to make public not
just our understanding of generally accepted principles of En-
glish but some of these private feelings and even private snob-
beries. I occasionally do so in this book—always, I hope, mak-
ing it apparent that that is what I am doing.
If good English were merely snobbery, it would still be worth
the attention of all except those who are immune to snobbery,
but it is more. There are positive reasons for valuing it. Al-
though readers may consult books like mine primarily to avert
criticism and save themselves embarrassment, in the long run
they are apt to find that they have also increased their pleasure
in using language and given others more pleasure in their use
of it.
In a sense, a language is an art form; in a sense, it is a game.
Those who appreciate or engage in painting or ballet are sen-
sitive to technique; so are those who appreciate or engage in
golf or tennis. Occasionally someone untrained in one of these
activities does something startlingly unconventional and won-
derful, just as a young child or a poorly educated or foreign-
born adult occasionally says something wonderful, makes
some truly creative use of English. A very few untrained practi-
tioners are even consistently remarkable—certainly this is true
in painting. Natural talent and something like luck play an
enormous role in art and in sports, and in language too. But
amateurs, no matter how talented or lucky they are, do not
generally experience or provide much pleasure at first—they do
not consistently please themselves or others. It is only as they
learn to respect conventions and techniques and begin to mas-
ter them that they reliably experience and provide pleasure.
Language is an artful game, sometimes casual and some-
times competitive, and those who know its conventions, tech-
niques, and finer points—those who have a command of good
English—play it better than those who don't. They are consis-
tent—and consistency, even in the details that are the subject
of Chapter 3 of this book, is an important secret of their game.
They can both please themselves and please others with their
play; they give their listeners or readers a good game. They also
win their way more frequently.
Good English is not the best English. The best English fre-
quently is good English, but the best users of English—the
great writers and poets, the great public speakers and con-
16. • Preface
versationalists—are often innovative and idiosyncratic and
therefore often less respectful of the strictures of good English
than most of us can dare to be if we want to avoid criticism.
Good English is more than merely adequate or serviceable,
however. It is English used well enough to give the user plea-
sure and to give pleasure to those who hear or read it, and if it
falls short of the beauty and grace of the best English, it still
reaches for beauty and grace and avoids the unbeautiful and
My définition of good English is as complete as I can make it
here—all the rest is in the details. I hope that those who use
this book and wrestle with its details not only will avoid
criticism but will find that the pleasures of language increase
for them and for those who listen to and read their words.
There remain a few comments about the organization and
coverage of the book and a suggestion on using it. Its four
chapters are a series of rules, each rule followed by examples
and explanations. The rules are for the most part the familiar
ones taught in primary and secondary schools, but the discus-
sions of them are extended unusually far—far enough to serve
sophisticated adult users of the language, those whose thought
is complex and whose verbal dilemmas are correspondingly
complex. Its coverage of punctuation and styling—that is, mat-
ters such as use of capitals and italics—is, I believe, more
comprehensive and more detailed than that of any other book
intended for general rather than professional use. It includes
some basic information on diction and composition.
The Glossary/Index at the back of the book defines and
illustrates grammatical terms and indexes the topics discussed
in the preceding four chapters. Extending its glossary function,
it also provides information and advice on many specific mat-
ters of English usage, in the manner of entirely alphabetical
handbooks, and thus it is quite long, unlike a typical glossary
or index. I have included these items, which in some cases are
brief versions of discussions in the preceding chapters and in
other cases concern specific words and details that are not
discussed or are discussed only glancingly elsewhere, so that
the book can have the handiness of an alphabetical guide as
well as the coherent structure of the topical guide it primarily
When the Glossary/Index does not answer the reader's ques-
tion directly or completely but refers to a rule, I advise reading
the entire discussion of the rule, even though some discussions
17. Preface •
are rather long. In such references I have often included the
wording of the appropriate subheading within the discussion,
which will make it easier to find the relevant passage, but
reading, or at least skimming, the entire discussion should
increase a reader's understanding of the general principles that
underlie the answer to a specific question and thus make
similar questions less troubling and less frequent in the future.
The book is intended to clarify general principles and hence
educate the reader, not just answer specific questions, though
it does that too.
19. THE

We learn the basic grammar of our native language, along
with its basic vocabulary, at a very early age and without
conscious effort. Then as we get older, the adults in our lives
become increasingly insistent that we learn correct grammar,
which seems to be made up of a lot of troublesome details that
must be learned consciously. When we get to school, we study
grammar more systematically and are exposed to special
terms—conjunction, gerund, predicate, and so on—used to
discuss it. We do learn quite a lot about grammar, but the
special terms give many of us difficulty, and almost all of us let
them fade from our minds when we leave school behind.
This chapter is concerned with correct grammar. It uses the
special terms, because there is no practical way to discuss
grammar without using them. However, when I introduce a
term that I think some readers may not understand, I define it
or give a simple defining example of it, and all grammatical
terms used in the book are explained in the Glossary/Index. A
reader who has unpleasant memories of struggling with these
terms as a child should find them quite easy to understand now
and may even get some pleasure from vanquishing gerund and
other bugbears of grammar school.
One grammatical term is grammar itself, and my use of it
requires some explanation. Throughout this chapter and this
book, when I state that something is ungrammatical or is
incorrect or faulty grammar, I am misusing the term grammar
22. • Grammar
as it is understood by scholars of language. To them, grammar
is not a set of rules that we should obey when using language
but a set of observations of how we do use language. If they
observe that many fluent native speakers of our language say
between you and I, they must conclude that English grammar
sometimes permits the preposition between to have the sub-
jective pronoun J as its object, though they may label the usage
in some way to indicate that it is not standard and is not in line
with broader observations about fluent use of English, such as
the observation that fluent users of the language generally use
the objective case, not the subjective case, for pronouns that
are the objects of prepositions.
This chapter, however, is not a scholarly study of grammar
but a guide to avoiding criticism for one's grammar. It assumes
that every reader's grammar is fluent, and in that sense correct.
Therefore I use the terms correct grammar and incorrect gram-
mar in their grammar-school senses: Correct grammar em-
ploys word relationships and form changes that are accepted as
correct by educators and the well-educated, and incorrect
grammar employs word relationships and form changes that
are condemned by them. Thus I call between you and I incor-
rect grammar, just as our schoolteachers did.
The rules and explanations in this chapter do not amount to
anything like a scholarly outline of English grammar. They are
merely intended to help fluent writers and speakers of English
avoid common errors—avoid faulty grammar—by making
them conscious of broad principles of English grammar that
they employ unconsciously whenever they use the language.
Principles that are understood only on a very deep mental level
are difficult to bring to bear on specific problems of expression
that we address consciously; we may suspect that something is
wrong but be unable to identify and correct the error unless we
can bring the principle involved to consciousness. In addition,
many errors in grammar do not violate deep principles at all—
they merely violate convention. Those who are not aware of
the principles and are therefore not aware of the difference
between a violation of principles and a violation of convention
must face every problem in expression in an almost super-
stitious way, hoping the jumble of half-remembered and quite
likely dubious precepts in their minds—Don't split infinitives;
Don't end a sentence with a preposition—will see them
The chapter includes some advice, such as on parallel con-
struction, that is concerned with effective use of language
23. The Sentence •
rather than strictly with grammar, because often it is the
choice we make among grammatical structures rather than
merely the Tightness or wrongness of those structures that
determines the overall quality of our expression. Conversely,
some matters that could be considered part of grammar are not
covered here but in other chapters—especially Chapter 2, on
punctuation, which reflects grammar and requires an under-
standing of grammar if it is to be used well—and in the Glos-
sary/Index. The Glossary/Index should be helpful to those who
want quick answers to specific questions. Sometimes it an-
swers a question directly, and sometimes it refers to the appro-
priate rule in this chapter or one of the others.
It is often difficult for those who do not know the name of
the error they may be committing to find the discussion of that
error in a reference book. I have done my best to reduce this
difficulty by careful listings in the Glossary/Index, but the
reader may have to do some skimming of the rules and their
discussions. To help the skimming eye, I have subdivided the
longer discussions, and when possible I have begun paragraphs
with examples of specific constructions that may match the
reader's problem.
Most of us don't have to be told what a sentence is. This is
fortunate, because it is possible to poke holes in any simple
definition. We can say that a sentence is a word group that
expresses a complete thought, but I said yes is a complete
sentence, yet hardly a complete thought; like many sentences,
it depends on its context to complete its meaning. We can say
that a sentence is a word group that includes a subject and a
verb, but Yes can be a complete sentence even though it has
neither subject nor verb, and When I came to dinner can't be a
sentence—at least out of context—even though it has both
subject and verb. Either the complete-thought definition or the
subject-and-verb definition could be expanded enough to make
it valid for just about all sentences, but we'd no longer have a
simple definition.
Since the subject of this chapter is grammar, we might try
the following definition: A sentence is a group of words that
are grammatically dependent on one another but are not gram-
matically dependent on any words outside the group. This
definition is not perfect, and it does not uniquely describe
24. 1-1 Grammar
sentences—it describes independent clauses too. However, it
does emphasize one important property of a sentence: the
grammatical dependence we expect the words within it to
Grammatical dependence is what determines whether a
group of words is a sentence, whether the group contains
enough words, too few, or too many, and whether the rela-
tionships among the words are easy or difficult for a listener or
reader to understand. The following five rules are concerned
with basic properties of good sentences—sentences that are
both good grammar and good uses of good grammar. (For a
discussion of types of sentences and clauses, see Rule 2-1.)
II 1-1 Write in whole sentences, not in
•I fragments.
/ discovered the overalls. When I was ladling out the chowder.
The fragment is easy to see. The second "sentence" is merely a
dependent clause of the first sentence. The word When makes
the clause dependent on something outside itself, so the word
group When I was ladling out the chowder does not meet the
definition proposed in the discussion just preceding this rule. It
must be joined to the first sentence, on which it depends: I
discovered the overalls when I was ladling out the chowder.
It may seem unlikely that a writer of any sophistication
would be guilty of fragments. Here is a more complicated
example: The President, whose term in office had hardly be-
gun when the opposition in Congress, which included mem-
bers of his own party, capitulated to public opinion, changing
the nature of his party leadership. The sentence is confusing,
and it takes some study to reveal that the confusion results
from a fragment. Was it the President or the opposition that
capitulated? If it was the opposition, then the whole sentence
is a fragment, because The President, which is obviously the
subject of the sentence, has no verb to be the subject of. If it
was the President that capitulated, then the opposition, just as
obviously intended to be the subject of a dependent clause, has
no verb, so the clause is a fragment.
Such fragments are common, particularly in journalism. A
hurried writer, or a hurried editor, may feel something is amiss
but not see the error—after all, it's hard to see what isn't there,
and often it's what isn't there that makes a sentence or clause a
25. The Sentence 1 -1
fragment. Whenever something seems wrong with a compli-
cated sentence, it helps to make sure that neither the sentence
as a whole nor any clause within it is a fragment.
A proper sentence generally contains a subject and a predi-
cate, but not every proper sentence does. And what of honor!
and So much for noble sentiments can stand alone as sen-
tences, though their meaning depends on the content of some
preceding sentence or group of sentences. They are not frag-
ments, because they are not grammatically dependent on any-
thing outside themselves and they do not require added words.
Fragments are sometimes deliberately employed to produce
special effects: / said a year ago that this company was headed
for trouble. Which is where we've arrived, as these figures will
show. There should ordinarily be a comma after trouble rather
than a period, but presenting the dependent clauses as if they
constituted a separate sentence gives them an emphasis that
may be desirable. The device should be used sparingly, and
alternatives should be considered; a dash after trouble would
give the clauses similar emphasis.
Sentences beginning with and or some other
And, but, or, for, so, yet, and other so-called coordinating
conjunctions are often used to begin sentences, despite an
older rule, still sometimes heard, that a sentence should never
begin with a conjunction because the conjunction makes the
sentence a fragment. It is true that a sentence that begins with
a conjunction—something joining its thought to the thought of
the preceding sentence—can hardly be anything but a fragment
of the complete thought, but that is no justification for such a
rule. After all, in a well-written paragraph each sentence
should add its thought to the thoughts of preceding sentences
whether or not it begins with a conjunction. Sentences that
begin with conjunctions are now accepted except in very for-
mal writing; I use them frequently in this book. To avoid them
we must either ( 1 ) actually connect the sentence to the preced-
ing sentence, which may be undesirable for a variety of rea-
sons; (2) replace the conjunction with a conjunctive adverb or
adverbial phrase (such as in addition for and, however for but,
alternatively for or, and consequently for so), which usually
also requires adding a comma after the adverb and may give
excessive emphasis to the connection to the preceding sen-
tence; (3) just drop the conjunction, which may remove a
26. 1 -2 Grammar
helpful indication of the significance of the statement to come;
or (4) completely recast the sentence.
It is acceptable to begin an occasional sentence with a con-
junction; such a sentence is not a fragment. But remember that
some people still condemn such use of conjunctions, and it can
lead to inept or confusing sentences (see also for in the Glos-
Elliptical sentences
Many sentences are elliptical—that is, they leave out one or
more words that the listener or reader can be expected to
supply. The missing word or phrase is called an ellipsis. An
elliptical sentence is not a fragment; fragments are faulty
grammar, but elliptical sentences are usually quite respectable
grammatically. (They are, however, sometimes ambiguous. For
example, John loves money more than Mary has an elliptical
dependent clause, which could be filled out in two very dif-
ferent ways: more than Mary loves money or more than he
loves Mary. See also Rule 1-3.)
Answers to questions are often elliptical. "When did you
discover the overalls!" "When I was ladling out the chowder."
In this dialogue, the answer is severely elliptical, leaving out
the entire main clause, which would be I discovered the over-
alls. But any listener or reader could supply the missing words;
the answer is still a complete sentence in its context. The
context can be more stately than conversation about Mrs.
Murphy's chowder: What is man! A featherless biped.
Il 1-2 Don't omit grammatically necessary
•I words.
The function of language is to communicate meaning, and
grammar is only one of the tools language employs to serve
that function. Yet meaning can be entirely clear and grammar
still faulty, just as meaning can be entirely clear in a sentence
with misspelled words. Good grammar has to be good in itself,
not just adequate to communicate meaning. Thus even when a
listener or reader would have no real trouble supplying an
omitted word, the omission may be an error if the word is
essential to the grammar of the sentence.
27. The Sentence 1 -2
Omission of parts of phrase pairs
The stock has always performed as well or better than ex-
pected attempts to be a compact sentence and does leave out
some dispensable words, but the second as in the adverbial
construction as well as should not be omitted; it should be as
well as or better than expected. The error is common in sen-
tences that include phrase pairs such as as well as . . . or better
than and as much as . . . if not more than. Thus The stock has
gone up as much if not more than IBM is a similar error. The
same errors occur with adjectival comparisons: Her money is
as green or greener than yours.
The stock has always performed as well as expected or
better and The stock has gone up as much as IBM if not more
are, however, correct. These are elliptical sentences (see Rule
1-1). It is permissible, and often desirable, to let the listener or
reader supply the missing words, which would be than ex-
pected in the first example and than IBM in the second exam-
ple. Thus though the first part of a phrase pair must be com-
plete, the second part can be elliptical. Ellipsis is part of the
language, and sometimes an essential part. Note that it occurs
elsewhere in these sentences as well. With every ellipsis filled,
the first sentence would be The stock has always performed as
well as it was expected to perform or better than it was ex-
pected to perform and the second sentence would be The stock
has gone up as much as IBM has gone up if not more than IBM
has gone up. Ellipsis saves us from such unnaturally tedious
Omission of words in compared items: false
Like the robbers, the cops' view of law enforcement is complex
omits too much, making a false comparison between the rob-
bers and the cops' view of law enforcement. It is two views, not
robbers and one view, that the sentence means to compare.
One way to repair the error is simply to make robbers an
independent possessive (see Rule 1-19), so that cops and rob-
bers share ownership of the phrase view of law enforcement:
Like the robbers', the cops' view of law enforcement is com-
plex. Another way would be to put the phrase in the first part of
the sentence and then repeat a word of it: Like the robbers'
view of law enforcement, the cops' view is complex. Still an-
other way would be a complete recasting: The cops, like the
robbers, have a complex view of law enforcement.
28. 1-2 Grammar
Profits were not so high as the preceding year and Profits
were higher than the preceding year make a false comparison
between Profits and the preceding year. Filled out, the sen-
tences would be Profits were not so high as they were in the
preceding year and Profits were higher than they were in the
preceding year. We can leave out they were—such an omission
is proper ellipsis (see Rule 1-1). And if we don't leave out they
were, we can even leave out in-, phrases such as in the preceding
year, which are called prepositional adverbial phrases, can
often be shortened by omitting the preposition, as in Quarterly
earnings will be announced [on] Friday. But we cannot leave
out both they were and in without creating a false comparison.
Since it is usually unlikely that such errors would mislead any
reader or listener, they are easy to make and to overlook; they
are somewhat disturbing, but it isn't immediately apparent
why. We all know that comparisons must be between items of
the same nature, and once we summon that very deep principle
to our conscious mind, the problem is quite apparent. Al-
though we can't write or speak fluently if our conscious mind
is cluttered with grammatical principles, we should be able to
bring these principles to consciousness when we need them.
Omission of verb forms
He either will or has already left is wrong. The verb form left is
appropriate with the second auxiliary verb, has, but inap-
propriate with the first, will. This kind of error is sometimes
called syllepsis. The sentence should be He either will leave or
has already left. Similarly, The country has already and will
continue going to the dogs is wrong; the verb form gone should
be supplied after already. If the form of a repeated verb changes,
it cannot be omitted in the first construction and supplied only
in the second. The verb can be omitted in the first construction
if it does not change form, as in He either is now or will soon be
leaving, in which leaving is the correct form in both con-
structions, but the omission may not always please the ear.
Changed verb forms can ©ften be omitted in the second
construction: / used the car when my father wasn't-, He didn't
go but should have-, He hasn't gone but will. When the first
application of the verb is omitted, it is an error of grammar, but
when the second application is omitted, it is a grammatically
permissible syllepsis, though it may be undesirable, as it is to
some degree in each of the three examples.
29. The Sentence 1 -2
When no auxiliary verb is involved but a verb changes form
because of a change in person, the verb can be omitted in the
second construction: / drive more than she-, I supply his finan-
cial support, his mother his emotional support. When an auxil-
iary verb is involved and changes form because of a change in
person, the whole compound verb can be omitted as long as the
form of the actual verb is the same, as in I am going to jail, you
to your just reward, in which the omitted auxiliary verb is are,
but the omitted actual verb is going, the same form as in the
first clause.
Sometimes an omitted verb has the same form as a supplied
verb but a different meaning. He is crazy already and quickly
driving his wife crazy may look fine—not only is the verb
supplied in the first construction but it is unchanged in form in
the second construction. However, the omission of is in the
second construction is at best questionable. In the first con-
struction, is is a linking verb—He is crazy—but in the second
construction, it is an auxiliary verb—He is . . . driving. The
same word should not be forced to carry two different mean-
ings, so is should be repeated in the second construction. Many
other verbs can have two or more distinct meanings—I have
gone, I have a gun-, He keeps fit, He keeps sheep, He keeps his
word—but is is the only one that is likely to be wrongly
omitted; no one would write He keeps fit, sheep, and his word.
Occasionally the multiple meanings of verbs are used deliber-
ately for a humorous effect, a device sometimes also called
syllepsis but more precisely called zeugma: He bolted the door
and his dinner-, He took his hat and his leave. See also zeugma
in the Glossary/Index.
You better do it right now is an odd but very common error;
the verb had is left out completely. In speech, You had better is
quite properly contracted to You'd better, then improperly
blurred to You better-, people come to consider it some sort of
idiom, or perhaps as the correct imperative You do it right now
with better thrown in as an intensifier, and use it even in
writing. It is incorrect in either speech or writing, though it
may eventually replace the correct form, and it is possible to
think up grammatical justifications for it (see better in the
Glossary/Index). Like any other error, it can legitimately ap-
pear in quoted dialogue, but I have seen it often in the dialogue
of fictional characters whom the writer did not mean to pre-
sent as careless speakers.
30. 1-2 Grammar
Omission of relative pronouns
He is the man went to Washington is distinctly folksy. How-
ever, He is the man we sent to Washington is good standard
grammar. We cannot ordinarily leave out a subjective relative
pronoun such as who, but we can often leave out an objective
relative pronoun such as whom. In simple sentences, the dis-
tinction is clear even with pronouns such as which and that,
which have the same form in subjective and objective cases; we
accept This is the house Jack built but not This is the house fell
down around Jack—we have to supply the pronoun which or
that to serve as the subject of fell. (When another clause inter-
rupts the relative clause, even a subjective relative pronoun is
sometimes omitted, as in This is the house I thought fell down
around Jack. See Rule 1-6 for more discussion of such inter-
rupting clauses.)
This is the house that Jack built and the weather destroyed,
leaving out that before the weather destroyed, is correct, and in
fact the first that can be omitted too: This is the house Jack
built and the weather destroyed. This is the house that col-
lapsed in the storm and fell down around Jack is also correct;
the single that can serve as the subject of both collapsed and
fell down. However, This is the house that Jack built and fell
down around him is incorrect. There must be a subject for fell
down around him, and the that earlier in the sentence will not
do, because it is already the object of the verb built. The same
relative pronoun cannot be used both as the object of one verb
and the subject of another, with the exception of the pronouns
whoever and whomever (see the discussion of pronouns as part
of their own clauses in Rule 1-6). In a complicated sentence, it
may take some study to reveal that a relative pronoun is trying
to play two grammatical roles. Thus They were all fully oc-
cupied in preparing for the invasion of the mainland, which
they had planned as the next stage in Allied strategy and was
to follow in less than a month is troubling—mysteriously so
until it is noticed that which is both the object of they had
planned and the subject of was to follow. But the error occurs
in simple sentences too, such as Do what you like and makes
you feel good, in which what is supplied as the object of like
but omitted as the subject of makes.
This is the house Jack built and that fell down around him is
correct, with that omitted as the object of built but supplied as
the subject of fell down. It is not, however, a pleasingly bal-
anced sentence; it would be much better with the objective
31. The Sentence 1 -2
that supplied. Some writers, as well as some editors, like to
omit every optional relative pronoun, but such a policy sug-
gests an excessively mechanical approach to language. An op-
tional pronoun often improves readability.
Note that in the examples above in which a relative pronoun
is correctly omitted, it always is part of a defining construction
rather than a parenthetical construction. A relative pronoun in
a parenthetical construction, such as which in This house,
which Jack built, fell down, can never be omitted, and it is
unlikely that any fluent user of English would omit it. For
discussions of defining and parenthetical constructions, see
the Glossary/Index and Rule 2-1.
Omission of a repeated preposition
We disagreed only with regard to what the disaster was due
has one too few uses of the preposition to, which is needed
after due as well as after regard: We disagreed only with regard
to what the disaster was due to. Similarly, It was a disaster the
significance of which no one was entirely ignorant needs of at
the end to go with ignorant-, the earlier of after significance
cannot play two roles.
It must be admitted that the correct versions of these sen-
tences are much harder on the ear or eye than the incorrect
versions, and that rewriting them would be advisable. Sen-
tences can end with prepositions, despite the oft-heard dogma
that they should not, but a sentence that does is likely to be a
sentence in which the word order is not standard, because in
standard word order a preposition is followed by its object.
Sometimes there is no good reason to depart from standard
word order. Certainly We disagreed only about the cause of the
disaster is easier and pleasanter to read than a sentence so
twisted that a preposition can be mislaid among its con-
Omission of a repeated modifier
There is enough time and energy, omitting the adjective
enough before the second object, is correct, but There is neither
enough time nor energy is faulty; it should be There is neither
enough time nor enough energy The error can be considered
faulty parallelism, which is discussed in Rule 1-5.
Body blows are the most reliable, effective, and punishing,
omitting the adverb most before the second and third adjec-
32. 1 -3 Grammar
tives, is correct, but Body blows are the most reliable, effec-
tive, and easiest to learn is faulty; since most does not apply to
easiest to learn, it should be supplied for effective. This error
too could be called faulty parallelism.
II 1-3 Don't omit words necessary to prevent
•I ambiguity or momentary misreading.
The preceding rule concerns omissions that leave meaning
intact but are grammatical errors. This rule concerns omis-
sions that are grammatically correct but produce ambiguity or
permit misreading.
John loves money more than Mary is ambiguous because the
than clause is elliptical. In most contexts the meaning would
be clear and the sentence might therefore be judged acceptable,
but in some contexts it might be unclear, and in any context it
could be criticized as imprecise. The than clause should be at
least partially filled out if precision is considered important:
than Mary does ox than he does Mary
He was expelled for failing physics and gambling is ambigu-
ous because of an omitted preposition; it should be He was
expelled for failing physics and for gambling, to prevent gam-
bling from being momentarily taken as a second direct object
of failing. Few readers would persist in their misreading and
believe that gambling was part of the curriculum. We uncon-
sciously and almost instantly correct such misapprehensions
when we read. Nevertheless they are annoying, and text that
contains many opportunities for misreading can be profoundly
irritating; somewhere below the level of consciousness, our
comprehension is continually backing out of blind alleys.
The word that is often omitted in such constructions as /
believe I'll go home and He said I could stay. These omissions
are fine, but sometimes when that is left out it is not clear
where it belongs. The expectation is falsely high earnings will
be reported could mean either The expectation is that falsely
high earnings will be reported or The expectation is falsely
high that earnings will be reported. Sentences with that omit-
ted should be inspected with extra care.
It takes special alertness to catch omissions that are gram-
matically correct but invite misreading, since we already know
what we mean. Ambiguity is always with us; the examples
above are merely a few of the many ambiguities that the En-
33. The Sentence 1 -4
glish language permits. Yet the effort to reduce ambiguity is
well worth making and should be part of the process of revising
any carefully composed work. See Rules 4-9 to 4-14 for advice
about that process.
II 1-4 Omit redundant or otherwise
II unnecessary words and phrases—
but with some discretion.
The traffic was as usual as ever is a typical careless redun-
dancy; as usual and as ever mean virtually the same thing.
This kind of redundancy repeats the same idea in different
words. It seems to be especially common with as con-
structions, as in Traffic was equally as bad last week-, either
equally bad or as bad should be used. The writer or speaker
may be using equally merely as an intensifier, like just, but to
the reader or listener, equally and as have the same meaning in
this context.
/ hope that when the parole board votes on my case that it
will not fail to consider my recent beatification incorrectly
repeats that. The first that introduces the remainder of the
sentence, which is a noun clause with an adverbial when
clause dependent on it. The second that reintroduces the noun
clause and should be taken out. The error is common when a
noun clause has a preceding dependent clause.
The examples above are true errors. More often, redundancy
is not an error but just an unnecessary use of a modifying word
or phrase. There are dozens of familiar expressions that cannot
be called grammatically incorrect but are redundant: con-
sensus of opinion means consensus-, variety of different
choices means variety of choices-, large in size means large-,
plans for the future means plans. One should watch out for
such redundant expressions—for one thing, they are overused
and consequently bore the reader, like clichés—but they do not
have to be exterminated; the cadence of a particular sentence
may make plans for the future more desirable than plans.
Writing from which every redundancy has been religiously
uprooted is apt to be unnaturally terse and clipped.
Refer back is often condemned as redundant, and it is redun-
dant in Please refer back to the previous chapter. But the re in
refer does not necessarily have the same meaning as back—
obviously it doesn't in Please refer to the next chapter. If I am
34. 1-5 Grammar
reading Chapter 10,1 might expect to be referred to Chapter 12
but would not object to being referred back to Chapter 8; the
back might be dispensable, but it would remind me that I am
being referred to text I have already read. It is wrongheaded and
simpleminded to leap on every redundancy.
Wordiness and flourishes
Because of the fact that I had occasion to be in possession of
the money, they were of the opinion that I was the party guilty
of having stolen it is wordy for Because I had the money they
thought I had stolen it. Such wordiness occasionally has a
function, emphasizing some part of the meaning or giving it a
slight twist, but usually wordiness suggests confusion, pom-
posity, or both. It is not an error of grammar but an error of
composition (see Rule 4-12). One might call it an overuse of
grammar—a use of complex grammatical structures to convey
a simple meaning.
/ venture to say that you wouldn't find me so contemptible if
I'd split the money with you begins with a somewhat quaint
flourish. However, an occasional flourish is not only permissi-
ble but desirable; flourishes can add nuance and expression to
otherwise bald statements and convey the feeling of the writer
or speaker about the statement. Of course, writers or speakers
who use / venture to say, I would hazard that, and similar
expressions to begin every other sentence—there seems to be
at least one such person at every conference table—are adding
off-flavor nuances; they are nervous, or pompous, or uncertain,
or just clumsy with language.
II 1-5 If there are elements in a sentence
II that are parallel in meaning and in
grammatical function, make them
parallel in grammatical form.
This is a basic rule of clear expression. Violations of the rule
are a feature of what one might call deliberately bad writing,
committed by writers who consciously vary the grammatical
form of parallel elements because they think the variation will
make their sentences interesting and impressive. Such varia-
tion may violate rules of grammar and will almost certainly
35. The Sentence 1 -5
make sentences needlessly confusing and clumsy. More often,
violations are accidental; writers merely fail to notice a poor
choice of phrasing, an omission of a necessary word (see Rule
1-2), or a mispositioning of a word. Correcting faulty paral-
lelism occupies more of an editor's attention than correcting
all other grammatical faults put together.
Items in a series not parallel
He liked sailing, swimming, and to fish is a simple example;
most of us don't have to be told that the third item in the series
should be fishing, producing a series of three gerunds rather
than two gerunds and an infinitive, or else the first two items
should be to sail and to swim, producing a series of three
infinitives. Yet wrong as the example seems, its grammar is
technically correct, since either a gerund or an infinitive can be
used as an object of liked. The error is an error of parallelism.
He liked sailing, beachcombing forays, and swimming is a
subtler example of faulty parallelism. Although sailing, beach-
combing, and swimming are all gerunds, beachcombing does
not stand alone but merely modifies the noun forays, so in-
stead of a series of three gerunds we have a gerund, a modified
noun, and another gerund. If we take out forays, the series is
properly parallel. The faulty parallelism in the example is only
faintly troubling, however, and one could even argue that it
gives the sentence a vitality that the stolid He liked sailing,
beachcombing, and swimming lacks. Rule 1-5 should not be
applied so zealously that every variation of structure in a series
is disallowed, especially in writing that is intended to do more
than merely state the facts.
Note that He liked sailing, swimming, and other seaside
activities is not a case of faulty parallelism. The third item in
the series is not parallel in meaning and significance to the
other two, but characterizes them and represents a group of
unnamed activities. Nor is He liked sailing, swimming, and
girls faulty parallelism; the series consists of two gerunds and a
noun, but there is no way to change the noun without changing
the content of the sentence—the series is as close to parallel as
it can be. Items in a series should usually be as parallel as their
meaning permits, but they don't have to be so parallel that we
can't say what we mean. He liked to sail, to swim, and girls is
faulty, however, because two infinitives and a noun combine in
a series much less happily—that is, they are farther from paral-
lel—than two gerunds and a noun.
36. 1 -5 Grammar
He liked to sail, swim, and to walk on the beach has a series
of three infinitives, but they aren't properly parallel. The word
to should either be eliminated before walk or be supplied
before swim. In putting to before the last infinitive but not the
middle one, the writer could be hoping to discourage a possible
but unlikely misreading; on the beach could grammatically go
with all three infinitives, as it does in He liked to sunbathe,
read, and sleep on the beach all day, though it would take a
perverse reader to notice the grammatical possibility in the
original example. If there is a real possibility of misreading
such a series, recasting to avoid the series is a better solution
than making the series nonparallel.
He liked to sail, swim, and had a passion for beachcombing
is in real trouble, because the last item is not part of the series
at all but is the second part of a compound predicate: He liked
. . . and had . . . The error seems glaring but is very common.
He liked to sail and swim and had a passion for beachcombing
is correct: two predicates to go with He, and two parallel
objects to go with liked. If we want to avoid the run-together
look of sail and swim and had, we can put a comma after swim
(a comma is usually unnecessary and undesirable between
compound predicates but is permissible to ease reading; see
Rule 2-3), or we can put in the comma and also repeat he before
the second predicate, making it an independent clause: He
liked to sail and swim, and he had a passion for beachcomb-
ing. See also false series in the Glossary/Index and the last
paragraph of Rule 2-6.
Either . . . or, not only . . . but also: correlative
items not parallel
Correlative items in a sentence are ones indicated by pairs of
conjunctions such as either . . . or, not only . . . but also, and
whether. . . or.
He has either gone swimming or someone has taken him
sailing is faulty parallelism—and faulty grammar—because the
second element is not a second predicate sharing the subject
He with the first predicate, but an independent clause with its
own subject, someone. The sentence can be made gram-
matically correct by changing the position of either: Either he
has gone swimming or someone has taken him sailing. Now
the correlative elements are both independent clauses. An-
other solution would be He has either gone swimming or been
taken sailing. Neither solution produces perfect parallelism—
37. The Sentence 1 -5
in the first, one verb is intransitive and the other transitive,
and in the second, one verb is active and the other passive.
However, both solutions are correct, and the parallelism can-
not be perfected without changing the meaning. For example,
He has either gone swimming or gone sailing loses the im-
plication that he can go swimming on his own but wouldn't be
expected to go sailing without someone else.
He has either gone swimming or gone sailing is precisely
parallel; gone swimming and gone sailing are grammatically
similar and share their relationship with he has. The sentence
can be made nonparallel all too easily by misplacing either: He
has either gone swimming or sailing omits a repetition of gone,
and He either has gone swimming or gone sailing omits a
repetition of has. These failures of parallelism are not really
offensive in the casual context of the example, but they are
noticeable. They could be considered uses of ellipsis (see Rule
1-1), but not every permissible ellipsis is a desirable one. The
sentence can also be made nonparallel by leaving either where
it was but repeating a word: He has either gone swimming or
has gone sailing unnecessarily repeats has. This failure of
parallelism is somewhat offensive; the ear and eye are more apt
to accept a questionable ellipsis than a questionable repetition.
The properly parallel sentence He has not only gone swim-
ming but gone sailing can be made nonparallel in the same
ways. With the conjunctive pairs either . . . or and not only. . .
but also, the item following the first conjunction and the item
following the second conjunction should be grammatically
Note that this is not true of all conjunctive pairs. With the
conjunctive pair whether. . . or, the item following the second
conjunction usually can be and often should be shorter. I don't
know whether he has gone swimming or he has gone sailing is
precisely parallel but not natural English; the second he should
come out, and has or has gone could come out.
He has either gone swimming or gone to town with his
father is not strictly parallel—gone swimming and gone to
town with his father are both predicates and hence are gram-
matically equivalent, but they are structured differently and
make different uses of the verb gone. That is quite all right;
correlative items should be as grammatically similar as their
meaning permits, but they cannot always be grammatically
identical. He has gone either swimming or to town with his
fathef is not all right; since gone functions differently with
swimming and to town, it should be repeated (see the discus-
38. 1 -5 Grammar
sion of omitted verb forms in Rule 1-2), and it can't be repeated
without repositioning either.
Sentences that are more ambitious than the examples above
often fall into misplacement of correlative conjunctions be-
cause of an inverted or otherwise unusual word order. The
effect is to make serious prose seem somewhat scatterbrained,
as in Not only had classical anticommunism returned to
Washington in official rhetoric, but also in military programs
and the reassertion of self-confidence. There is a failure of
parallelism, because the item introduced by Not only is a
clause, but the item introduced by but also is merely a preposi-
tional phrase. The latter item could be made a clause, of
course: . . . but it had also returned in . . . Parallelism could
also be achieved by using either standard word order—Classi-
cal anticommunism had returned to Washington not only in
official rhetoric but also in military programs and the reasser-
tion of self-confidence—or a different nonstandard order—Not
only in official rhetoric but also in military programs and the
reassertion of self-confidence had classical anticommunism
returned to Washington. See also the discussion of complica-
tions in inverted sentences in Rule 2-5.
More than, as much as: adverbial comparisons not
Adverbial comparisons in a sentence are ones joined by phrases
such as more than and as much as. Errors occur with them
(and with adjectival comparisons, such as greener than and as
green as) when a necessary than or as is omitted, as discussed
in Rule 1-2. Errors also occur when the second item in the
comparison is a pronoun, as in He sails more than me, which
can be considered an error of parallelism, since He and me aie
grammatically parallel and should therefore be in the same
case (such errors are discussed in Rule 1-6 as errors in case).
He didn't like swimming as much as to sail is clearly non-
parallel and ugly. However, lack of parallelism can be much less
apparent in more complicated sentences, and it can be defensi-
ble. He learned to swim that summer, but more than swim-
ming with his friends on the broad public beach he liked to
sail to the deserted strands of the islands in the bay fails to
make swimming and to sail parallel, but then perhaps they are
not really parallel in thought anyway—there is an implication
that when he got to those deserted strands he liked to swim
there, and consequently the parallel in thought is between
39. The Sentence 1 -5
swimming with friends and swimming alone rather than be-
tween swimming and sailing. English is not mathematics, and
language can sometimes compare nonparallel things—can
compare apples and oranges. Careful parallelism is not the only
important property of good English, and sometimes it is a
dispensable property.
But not, rather than: antithetical constructions not
Antithetical constructions are used to state that something is
true of one thing but untrue of another. He liked sailing and
swimming but not to walk on the beach is faulty parallelism;
to walk should be changed to walking. When the untrue item
is given first, but not becomes not . . . but, and errors of
parallelism can occur in the same way they do in correlative
constructions, discussed earlier in this rule: He has not gone
swimming but sailing omits a desirable repetition of gone, He
has gone not swimming but gone sailing undesirably repeats
gone, and so on.
He chose to sail to the island rather than swimming there is
nonparallel, and it is easily made parallel by changing swim-
ming to to swim or simply to swim—it is often permissible to
leave out to in an infinitive, though to should be either consis-
tently included or consistently omitted in the second and sub-
sequent infinitives in a series, as explained earlier in this rule.
However, nonparallelisms with rather than are often not objec-
tionable, even in such a straightforward sentence as the exam-
ple, and sometimes they are necessary. He sailed to the island
rather than swam there is parallel, and He sailed to the island
rather than swimming there and He sailed to the island rather
than swim there are not, but the second and third versions do
not mean the same as the first; the first version simply tells us
what he did and did not do, whereas the second suggests to us
and the third tells us that he made a conscious decision be-
tween alternatives. When the negative rather than con-
struction precedes the positive construction, parallelism is
actually an error: Rather than swam there, he sailed to the
island is not English, though the nonparallel swim and swim-
ming would both be English. The normally conjunctive phrase
rather than is often used, and used correctly, as if it were a
prepositional phrase such as instead of, and when it is so used,
the rule that items joined by conjunctions should be as gram-
matically similar as possible must sometimes be abandoned.
40. 1-5 Grammar
Like and unlike
These words very often occur in introductory constructions:
Like me, she is a teaching fellow-, Unlike her classwork, her
tutorial duties bore her. They seem to invite faulty parallelism,
and the result is false comparison (discussed in the Glossary/
Index and in Rule 1-2). Like me, tutorial duties take up a lot of
her time and Unlike her classwork, she is bored by her tutorial
duties are examples; in the first, me is not parallel to tutorial
duties, and in the second, classwork is not parallel to she. The
frequency of such errors may be partly due to haziness on the
proper functions of like-, see also like for as, as if, or as though
in the Glossary/Index.
The case of a noun or pronoun is determined by the function of
the word within its sentence—by whether it is the subject of a
verb, the object of a verb or preposition, or the possessive
modifier of another word. English nouns have only two forms
for the three cases, since the subjective and objective forms are
the same; the possessive case is formed by adding an apos-
trophe and s or sometimes just the apostrophe (see Rule 2-29).
Some pronouns, such as one and anybody, also have only two
forms, but some others have not just three but four. I, me, and
my are subjective, objective, and possessive forms, and there is
also a special form for the so-called independent possessive,
mine, which instead of merely modifying another word acts
like a noun: Let's take your car, since mine has bald tires. The
possessive mine can even itself be made possessive—Let's take
your car; mine's tires are bald—though this is not true of other
independent possessives, such as yours and theirs.
Except for independent possessives, possessive nouns and
pronouns are actually modifiers, and they are discussed later in
this book (Rule 1-19), though Rule 1-7 concerns the use of the
possessive case for the subject of a gerund. The three other
rules in this section concern the pronouns that have different
forms for the subjective and objective cases.
41. Case of Nouns and Pronouns 1 -6
Il 1-6 Put the subject of a verb in the
•' subjective case.
Since nouns have the same form in the subjective and objective
cases, violations of this rule occur only with a few pronouns—
the personal pronouns I/me, he/him, she/her, we/us, and they/
them, and the relative or interrogative pronoun who/whom
and its indefinite form whoever/whomever. But because these
pronouns are common, errors in case are common.
Pronouns as part of compound subjects
Johnny and me want to go swimming is amazingly difficult to
stamp out of a child's speech. This may be one of the times that
natural grammar—the grammar we absorb as we learn to speak
and long before we go to school—is at real odds with standard
English. The child perhaps considers Johnny and me, or even
me and Johnny, to be a single idea that should keep the same
form whether subject or object. Eventually, parents and teach-
ers convince the child that the pronoun in a compound subject
has to have the same form that it would if it were standing
alone—I want to go swimming—and we begin to hear John and
I want to buy a motorcycle.
Pronouns as part of their own clauses
I avoid him who has the plague is correct; him is the object of
avoid, and who is the subject of has, the verb in its own clause.
Those who make it I avoid he who has the plague may just be
afraid of the objective case, having in childhood been corrected
so often about Johnny and me want to go swimming, but more
sophisticated people make the error too, because it does seem
to have some logic going for it. The entire word group him who
has the plague acts as a unit—in the example, as the direct
object of avoid—and the who in the subordinate clause seems
to attract the him of the main clause to its own case. It
shouldn't; who is governed by its own clause, in which it is the
subject, but him is governed by the main clause, in which it is
the object.
It was she I was writing about may seem puzzling at a
glance, because there is no objective pronoun and J was writing
about seems to require one. The temptation is to make it It
was her I was writing about, thus providing an objective pro-
42. 1 -6 Grammar
noun. This is an error—It was she is correct, since a pronoun is
governed by its own clause. The missing objective pronoun,
whom, has simply been omitted, as is entirely permissible (see
the discussion of omission of relative pronouns in Rule 1-2).
With the ellipsis filled in, the sentence becomes It was she
whom I was writing about.
I invited people whom I thought would get along together is
just as wrong, if less apparently so, as / invited people whom
would get along together. The pronoun whom is the subject of
would, not the object of thought, and it should therefore be
who. Often a relative clause such as who would get along
together is interrupted by another clause such as J thought.
The object of the verb in the interrupting clause is somewhat
difficult to pin down. In effect it is the idea, but not the exact
words, of the surrounding clause: J thought they would get
along together. Perhaps this fuzziness about the object of the
verb in the interrupting clause explains an odd fact. Even
though the relative pronoun in the sentence I invited people
who I thought would get along together is subjective, it can be
dropped: / invited people I thought would get along together.
Normally we could not omit a subjective pronoun—we could
not make it I invited people would get along together—but the
interrupting clause permits the omission, just as if the pro-
noun were objective, as it is in I invited people whom I thought
you would like.
Whom shall I say is calling! is a common error among those
who think whom is always more genteel than who (see gen-
teelism and hyperurbanism in the Glossary/Index). More so-
phisticated people make the error too, particularly in passive
constructions, such as Whom did you say was being invited!
Here the interrupting did you say camouflages the otherwise
glaring wrongness of Whom was being invited!
I saw a man who I thought was better dressed than I and /
met a man whom I thought to be better dressed than I are both
correct. In the second example, whom is objective as the sub-
ject of the infinitive to be (see infinitive in the Glossary/Index).
In I saw a man whom I thought better dressed than I, the
infinitive is omitted but understood, and whom remains cor-
rect. (Rule 1-6 would be more precise if I had made it "Put the
subject of a finite verb in the subjective case"—but I did not
want to puzzle readers with the term finite. Seefiniteverb in
the Glossary/Index.)
Whoever, unlike other pronouns, can play two roles in a
sentence at once. It can function as the subject of one verb and
43. Case of Nouns and Pronouns 1 -6
the object of another, as in / will invite whoever wants to
come, in which whoever is the subject of wants and also the
object of invite (though more precisely it is not whoever but
the entire clause whoever wants to come that is the object of
invite). Whomever can function as the object of verbs in two
clauses, as in / will invite whomever you choose, or as the
subject of a verb and the object of a preposition, as in
Whomever we send invitations to is sure to come and For
whoever draws the lucky number there will be a prize. Other
combinations of function are possible. As the examples here
show, the form of the pronoun—whether it is the subjective
whoever or the objective whomever—is determined by the role
it plays in its own clause, which is the clause that completes its
meaning, defining who whoever or whomever is. In speech,
occasional errors are almost inevitable, because the role of the
pronoun can be so complicated, as in This invitation is for
whoever that is you're with—the temptation is strong to make
it whomever, as the object of for or the object of with. When we
are writing, we have time to figure out that whomever that is
would be an error. See also Rule 1-8 and who, whom-, whoever,
whomever in the Glossary/Index.
Pronouns in elliptical clauses
She sails better than him seems wrong to most of us, and to all
of us if the elliptical clause is filled in: She sails better than
him sails. The word than is a conjunction, and conjunctions
join words or word groups of similar grammatical signifi-
cance—two adjectives modifying the same noun, two subjects
or two objects of the same verb, two clauses, and so on. In She
sails better than him, than joins a clause and an objective
pronoun, which is not a proper function of a conjunction. Use
of objective pronouns with than has been exceedingly common
for centuries, however, especially with first-person pronouns:
She thinks she's better than me-, She sails better than us.
Consequently, some modern dictionaries accept than as a prep-
osition, condoning its use with objective pronouns, since the
objects of prepositions should be in the objective case (see Rule
1-9). I advise denying oneself this liberty, since there are many
who condemn it.
She likes him better than me is a correct use of than as a
conjunction. Me is objective, but that is all right, because it is
the objective pronoun him that me is joined with, and if the
elliptical clause is filled in, the sentence becomes She likes
44. 1-6 Grammar
him better than she likes me. When in doubt about the proper
case for a pronoun following than, we can just imagine the
sentence with the elliptical clausefilledin. See also than in the
He has better friends than I is correct but ambiguous; it
could mean either He has better friends than I am or He has
better friends than I have. Elliptical clauses should be checked
for ambiguity as well as for grammatical soundness.
Pronouns in apposition
John, he of the big mouth, won't be invited and Let's not invite
John, him of the big mouth are both correct. In the first sen-
tence, he of the big mouth is in apposition to John, the subject
of the sentence, and the pronoun he is in the subjective case. In
the second sentence, him of the big mouth is again in apposi-
tion to John, but John is the object of the sentence, and the
pronoun him is in the objective case. The case of a pronoun in
apposition is determined by the case of the word that it is in
apposition to. (See also Rule 1-19 for special problems with
The directors you have chosen, Mr. Smith and me, will do
our best is an error; Mr. Smith and me is in apposition to The
directors, the subject of the sentence, and hence should be Mr.
Smith and I. The intervening you have chosen encourages the
error—its understood object, the relative pronoun whom,
seems to offer its invisible self for Mr. Smith and me to be in
apposition to.
All of us are going may seem puzzling, since Us are going is
impossible. But in All of us are going, the pronoun us is the
object of the preposition of, not a word in apposition to the
subject of the verb; there is no apposition in the sentence. The
entire phrase All of us is the subject, and the case of the
pronoun is determined by its role within its phrase.
T. S. Eliot's line Let us go then, you and I could be considered
an error; you and I is in apposition to us, and thus it should be
you and me. However, you and I is supported by idiom and to
some extent by grammatical analysis. Let us and Let's are so
frequently followed by subjective pronouns that objective pro-
nouns are apt to seem wrong or at least colloquial, as in the
correct Let's you and me have a drink. Let us and particularly
its contraction Let's are not perceived as what most gram-
marians say they are, the imperative Let and the objective us or
its contraction. One of the most scholarly grammarians,
45. Case of Nouns and Pronouns 1 -6
George O. Curme, would consider Let us go to be a subjunctive
rather than an imperative construction, a modern form of Go
we, and Go we then, you and I could not be attacked for
disagreement in case, so perhaps Let us go then, you and I
should not be attacked either. Nevertheless, I advise not using
the subjective after Let us and Let's, if only because Let's you
and I has at least a faint whiff of the reeking gentility of
between you and L, people who use the subjective may be
suspected of doing so not because they tolerantly accept idiom
but because they intolerantly and ignorantly think the subjec-
tive is more elegant.
Let's encourages other apposition errors besides errors of
case, such as the colloquial Let's us go and Let's you and him
make up, which when the contraction is expanded become the
grossly redundant Let us us go and the nonsensical Let us you
and him make up. Obviously, let's has acquired a broader
meaning than that of the uncontracted let us, but in anything
more than casual speech it should not be used where let us
cannot be used.
Its me or It's /? Pronouns as subject complements
A subject complement is a word or phrase that follows a link-
ing verb such as is or seems} it's the that in This is that, and it's
the gray in All cats seem gray A subject complement isn't the
object of a verb but something linked to the subject by a verb.
The rule for subject complements is very simple: They should
be in the same case as the subject they are linked to, which is,
of course, the subjective case.
It's me and It's us break the rule, a fact that has probably
generated more incredulity among grammar-school students
than any other precept of "good grammar," because It's I and
It's we seem impossibly unnatural to them. I advise breaking
the rule whenever the subjective pronouns / and we seem stiff
or prissy, as they do following the informal contraction It's and
in many other situations. That was we singing outside your
window last night-, When you hear three knocks, it will be I-,
His chief victim was I—such sentences may obey the rule, but
they are idiomatically objectionable. There are, of course, sen-
tences in which obeying the rule is not idiomatically objec-
tionable. In It was I who broke your window, the subjective
who seems to make I preferable even though in principle there
need be no agreement in case between a pronoun and its ante-
cedent (see Rule 1-12). The ear has to be the judge.
46. 1-7 Grammar
It's him and It's her cannot be defended quite as energetically,
because the rule-observing It's he and It's she, though perhaps
slightly stilted, are not outlandish; most careful speakers and
writers do use them. It's them is perhaps more often defensi-
ble, because It's they is more than slightly stilted. Again, the
ear must be the judge; That was he singing outside your win-
dow seems fine to me, but His chief victim was she seems
contrary to idiom, and to a lesser extent so does That was they
singing outside your window.
II 1-7 Put the subject of a gerund in the
II possessive case, if possible.
I dislike that man's wearing a mask and / dislike that man
wearing a mask are different statements. In the first, the wear-
ing of the mask is disliked; in the second, the man is disliked.
In the first statement, wearing is a gerund—that is, a special
verb form that functions as a noun—and it is the object of the
sentence, with the possessive phrase that man's modifying it.
Such a possessive "owns" the action implied by the gerund and
thus is considered the subject of the gerund. In the second
statement, wearing is a participle—that is, a special verb form
that functions as an adjective—and that man is the object of
the sentence, with the participial phrase wearing a mask modi-
fying it.
However, very often the objective case rather than the pos-
sessive case is used for the subject of a gerund, especially when
it is unlikely that the gerund will be misperceived as a partici-
ple, as in / dislike him wearing a mask. Many writers and
editors, and some of the grammarians whose books they use for
reference, consider use of the objective case for the subject of a
gerund to be standard idiomatic English, and certainly it is
common. Other writers and editors, and the grammarians they
prefer, condemn use of the objective case if the possessive case
is possible. Since such use of the objective case will not escape
criticism, I advise against it. I also believe that it eliminates a
useful grammatical signal and permits an annoying fuzziness
of syntax. A sharper understanding of what a gerund is may
help reduce the fuzziness.
There are two types of gerund. One type is exactly like a
noun—it can be the subject or object of a verb, it is modified by
47. Case of Nouns and Pronouns 1 -7
articles and adjectives, and it cannot take a direct object. The
other type is mostly like a noun but has some of the charac-
teristics of a verb or a participle—it too can be the subject or
object of a sentence, but it is modified by adverbs and can take
a direct object. In The inappropriate wearing of a mask is
forbidden, the gerund wearing is of the first type; in Inap-
propriately wearing a mask is forbidden, the same gerund is of
the second type. Of course, a gerund with no modifier and no
object or of phrase following it cannot be assigned to either
type. We do not mix the types in modern English, though
fluent users of the language did mix them in previous cen-
turies. The journals of the eighteenth-century explorer James
Cook are full of examples, such as The trouble and vexation
that attended the bringing these animals thus far is hardly to
be conceived, in which bringing is modified by the, just as a
noun would be, but has the direct object these animals, just as
a verb or participle would have.
Every modern fluent user of English automatically uses the
possessive for the subject of gerunds of the first type—/ dislike
that man's inappropriate wearing of a mask—because the
"nounness" of the gerund is so evident. But a great many fluent
speakers and writers use the objective for the subject of
gerunds of the second type—J dislike that man inappropriately
wearing a mask—because the "nounness" of the gerund is
obscured by its adverbial modifier and direct object. When the
objective is used instead of the possessive, the gerund can be
perceived as a participle modifying man rather than a gerund
modified by man, and the meaning is likely to be different.
Sometimes it makes little difft rence to the sense of a sentence
whether a verb form ending in ing is understood as a participle
or as a gerund. For example, / don't remember his ever being
angry and / don't remember him ever being angry mean very
nearly the same thing, and an argument could be made for
preferring the latter—the thought is of the man angry more
than of the man's anger. But often there is a difference, and if
we mean the ing word to be a gerund rather than a participle,
we should use the possessive case for its subject.
She approves of the teacher handing out extra homework as
punishment would probably not be misunderstood; almost
certainly the approval is of the handing out of the homework,
not of the teacher observed to be handing it out. But a usage
that is unlikely to be misunderstood is not necessarily a usage
that should be accepted as correct. At least in principle the
example is just as wrong as She approves of the teacher disci-
48. 1 -7 Grammar
pline, in which the gerund phrase has been replaced by a noun.
The subject of a gerund "owns" the action of the gerund, and
owning is expressed by the possessive case. She approves of the
teacher's handing out extra homework as punishment is there-
fore preferable.
When the possessive is impossible or bizarre
When the subject of a gerund is not a simple noun or pronoun
but a group of words, it may be impossible or at least bizarre to
use the possessive. For example, the plural in Many of us don't
approve of a man whom we voted against's being elected is
bizarre. It may seem reasonable enough to dispense with the
possessive in such situations: . . . a man whom we voted
against being elected. However, we would not write Many of
us don't approve of a man whom we voted against's election
either, and we would not have the alternative of dispensing
with the possessive; . . . a man whom we voted against elec-
tion is not English. We would rephrase, using an of con-
struction: Many of us don't approve of the election of a man
whom we voted against. We are not forced to rephrase with the
gerund as we are with the noun—but we could choose to
rephrase. We accept a man whom we voted against being
elected only because the objective rather than the possessive is
so often used for the subject of a gerund even when the pos-
sessive is not impossible; the objective never surprises us. But
its failure to surprise us does not make it desirable. I advise not
accepting it without some thought; rephrasing to avoid it may
be worth the trouble. Sentences in which the possessive is
logically called for but is impossible are likely to be clumsy
Many of us don't approve of this man, whom we voted
against, being elected can have no possessive for the subject of
the gerund, not even a bizarre one, because the relative pro-
noun whom cannot have a possessive form as its antecedent
(see Rule 1-19). But rephrasing remains an option.
Sometimes rephrasing is not a good option and it is wiser to
accept the objective subject of the gerund. There is no law
against gambling, but there is a law against people actively
involved in a sport betting against themselves might be such a
case. Some words, such as any, never or only very rarely have
possessive forms, so we use the only form available: Some
players have been hurt, but I've never heard of any dying.
English does accept the objective for the subject of a gerund
49. Case of Nouns and Pronouns 1 -8
when there is no reasonable alternative. Yet usually there is an
alternative phrasing. For example, There is no sense in both of
us going cannot be called an error—it is virtually an idiom, and
certainly both of us cannot be made possessive. The fastidious
may nevertheless make it There is no sense in our both going,
which is just as idiomatic and allows the possessive.
Confusion of gerunds with participles in absolute
John having worn a mask, no one knew he was there begins
with an absolute construction (see Rule 1-21 and absolute
construction in the Glossary/Index). The word having is not a
gerund but a participle. Past participles can be used in absolute
constructions too: The mask removed, we all recognized John.
John's having worn a mask, no one knew he was there is a
bad error. The possessive should be used for the subject of a
gerund, but not for the subject of a participle—that is, for the
word the participle modifies. The error is infrequent, but
someone trying hard to follow my advice and use the pos-
sessive with gerunds might slip into it.
I found a fine example of a gerund construction in an Amer-
ican grammar published in 1863: Caesar's having crossed the
Rubicon spread consternation throughout Rome. The gerund
construction can be made an absolute construction by chang-
ing Caesar's to the subjective case, inserting a comma, and
making consternation the subject of the basic sentence: Caesar
having crossed the Rubicon, consternation spread throughout
II 1-8 Put the object or indirect object of a
II verb or verbal in the objective case.
Like Rule 1-6, this rule is violated only with the few pronouns
that have different forms for the subjective and objective cases.
Pronouns as part of compound objects
Our parents sent John and I to Europe and Our parents gave
John and I a trip to Europe are embarrassing errors, much
worse than the childish Johnny and me want to go swimming.
Not only are they incorrect, they also suggest a self-conscious
50. 1-8 Grammar
effort to be correct—they are hypercorrect (see hyperurbanism
in the Glossary/Index). Once Johnny and me want to go swim-
ming is eradicated, some of us go too far and give up the
objective case in compound objects, though very few of us
would fail to use the objective case for a pronoun standing
alone as object—sent I to Europe and gave I a trip are quite
evidently not English.
Pronouns as part of their own clauses
/ avoid he who has the plague is incorrect, because the pro-
noun he is the object of avoid—the verb in its own clause—and
should be him. See Rule 1-6 for more discussion of this point.
Pronouns as objects of verbals
/ hate saluting him-, I hate to salute him-, The man saluting
him must be his son. The objects of verbals—that is, of
gerunds, infinitives, and participles—are always in the objec-
tive case. The subjects of verbals are not so consistent—the
subject of a gerund should usually be in the possessive case
(Rule 1-7), the subject of an infinitive should be in the objective
case, as in / want him to salute me, and the subject of a
participle can be either subjective or objective, depending on
its role in the sentence. But the objects of verbals are always
A problem: who and whom, whoever and whomever
Who is the subjective case and whom is the objective case, and
we can, if we like, apply Rule 1-8 strictly: Whom are you going
to invite! Whom are you going to send invitations! But for a
century and a half, language arbiters from Noah Webster on
have been pointing out that educated speakers and writers
often use who and whoever when the objective case is called
for: Who are you going to invite! I'm going to invite whoever I
choose. Certain failures to use the objective are perceived as
glaringly wrong, such as To who will you send invitations! But
most get by, and their correct equivalents can seem labored and
prissy. For some reason, whom and whomever have always had
a la-di-da flavor.
In formal writing it is best to follow Rule 1-8 strictly and use
whom and whomever in every objective situation. In less for-
51. Case of Nouns and Pronouns 1 -9
mal writing and in speech it is permissible to use who and
whoever when they seem more natural. This way we are at
least less likely to make the foolish error of using whom when
it should be who and ending up both la-di-da and wrong.
Whoever can be, or at least seem to be, both the subject of
one verb and the object of another, as in I'm going to invite
whoever wants to come and Whomever you invite is likely to
refuse. Its case is determined by the role it plays in its own
clause—the clause that explains who whoever is. See also Rule
II 1-9 Put the object of a preposition in the
I' objective case.
She wrote the most lovely note to John and I and I don't
understand what's going on between Mary and he are embar-
rassing errors, because they suggest an attempt to be elegant.
Such errors with objects of prepositions are quite common; for
some reason, people who would not break Rule 1-8 by saying or
writing Mary drove John and I home or Mary gave John and I a
lift will break this rule, sometimes even when a pronoun
directly follows a preposition: to he and I, between she and he.
The object of to, between, or any other preposition must be in
the objective case, just as the object or indirect object of a verb
must be.
Mistakes with who are comparatively rare; few people say or
write For who are you going to votel If the preposition does not
immediately precede the pronoun it is permissible, as ex-
plained in Rule 1-8, to use who instead of whom: Who are you
going to vote fori
Everyone but he left and Everyone left but I are errors. In
many other constructions, the versatile word but is not a
preposition; for example, in She left but he didn't it is a con-
junction. But in Everyone but he left and Everyone left but I, it
is preposition, with the same meaning as the preposition ex-
cept, and its object must be in the objective case: Everyone but
him left-, Everyone left but me. Note that Everyone left but he
cannot be passed off as elliptical for Everyone left but he didn't
leave, which is a logical contradiction; if he didn't leave, then
it is false to say that everyone left.
Don't act like I'm going to bite you is a very common error.
52. 1-10 Grammar
Since like is a preposition, the example does violate Rule 1-9,
but the error is usually committed not because of ignorance of
the rule but because of a misunderstanding of the word like,
which should not be used to mean as or as if, which are
conjunctions. See also like in the Glossary/Index.
The preposition of is sometimes followed by the possessive
case, as in Any friend of John's is a friend of mine, in which
both John and mine are possessive. See possessive case in the
Long before the schoolteachers get hold of us, we learn that in
an English sentence certain words must agree in form with
certain other words—that He don't and John and Mary is in
love are faulty grammar. A verb's form may be affected by
whether its subject is in the first person [I), the second person
(you), or the third person {he, she) and by whether the subject is
singular or plural, and we pick up this part of grammar as we
learn to talk.
Applying the principles of agreement is not really very diffi-
cult, since English, unlike many other languages, does not have
separate inflections, or form changes, for every situation.
Nevertheless, errors do occur, and disagreement in number,
covered in Rules 1-11 and 1-12, is common even in simple
situations. Also, the very simplicity of English inflection can
be a problem, since a word may be in grammatical agreement
with too many other words in the sentence, permitting ambi-
guity; avoiding such ambiguity is the concern of Rule 1-13.
II 1 - 1 0 Make a subject and its verb agree In
II person.
You are crazy and / am not crazy are straightforward examples
of subject and verb agreement in person. But should it be Either
you or I is crazy, or am crazy, or are crazy2. The verb has two
subjects but can agree with only one—a situation called syllep-
sis (see the discussion of omission of verb forms in Rule 1-2).
Syllepsis is sometimes an error, but in either . . . or con-