The King’s English

PREFACE

PREFACE TO THE SECOND EDITION

CONTENTS

CHAPTER I VOCABULARY — General — Malaprops — Neologisms — Americanisms — Foreign Words — False, Ugly, or Needless Formations — Slang — Particular Words — FOOTNOTES:

CHAPTER II SYNTAX — Case — Number — Comparatives and Superlatives — Relatives — Participle and Gerund — Participles — The Gerund — Shall and Will — The Perfect Infinitive — Conditionals — ‘Doubt that’ and ‘doubt whether’ — Prepositions — FOOTNOTES:

There is not much opportunity in English for going wrong here, because we have shed most of our cases. The personal pronouns, and who and its compounds, are the only words that visibly retain three—called subjective, objective, possessive. In nouns the first two are indistinguishable, and are called the common case. One result of this simplicity is that, the sense of case being almost lost, the few mistakes that can be made are made often—some of them so often that they are now almost right by prescription.

1. In apposition.

A pronoun appended to a noun, and in the same relation to the rest of the sentence, should be in the same case. Disregard of this is a bad blunder.

2. The complement with am, are, is, &c., should be subjective.

In the last but one, him would no doubt have been defended by the writer, since the full form would be he whom, as an attraction to the vanished whom. But such attraction is not right; if he alone is felt to be uncomfortable, whom should not be omitted; or, in this exalted context, it might be he that.

On that’s him, see 4, below.

3. When a verb or preposition governs two pronouns united by and, &c., the second is apt to go wrong—a bad blunder. Between you and I is often heard in talk; and, in literature:

4. The interrogative who is often used for whom, as, Who did you see? A distinction should here be made between conversation, written or spoken, and formal writing. Many educated people feel that in saying It is I, Whom do you mean? instead of It’s me, Who do you mean? they will be talking like a book, and they justifiably prefer geniality to grammar. But in print, unless it is dialogue, the correct forms are advisable.

5. Even with words that have no visible distinction between subjective and objective case, it is possible to go wrong; for the case can always be inferred, though not seen. Consequently a word should never be so placed that it must be taken twice, once as subject and once as object. This is so common a blunder that it will be well to give a good number of examples. It occurs especially with the relative, from its early position in the sentence; but, as the first two examples show, it may result from the exceptional placing of other words also. The mere repetition of the relative, or insertion of it or other pronoun, generally mends the sentence; in the first example, change should only be to only to be.

Analogous to these are the next three examples, which will require separate comment:

To is applicable to add, not to take away. The full form is given by substituting for or ‘and from the certainty of which no authority could’. This is clearly too cumbrous. Inserting or from after to is the simplest correction; but the result is rather formal. Better, perhaps, ‘the certainty of which could not be increased or diminished one jot by any authority’.

A second in is required. This common slovenliness results from the modern superstition against putting a preposition at the end. The particular sentence may, however, be mended otherwise than by inserting in, if excel is made absolute by a comma placed after it. Even then, the in would perhaps be better at the end of the clause than at the beginning.

The italicized upon (we have nothing to do with the other upon) is right with insist, but wrong, though it must necessarily be supplied again, with impress. It is the result of the same superstition. Mend either by writing upon after insisted instead of before which, or by inserting which he after and.

6. After as and than.

These are properly conjunctions and ‘take the same case after them as before’. But those words must be rightly understood. (a), I love you more than him, means something different from (b), I love you more than he. It must be borne in mind that the ‘case before’ is that of the word that is compared with the ‘case after’, and not necessarily that of the word actually next before in position. In (a) you is compared with him: in (b) I (not you) is compared with he. The correct usage is therefore important, and the tendency illustrated in the following examples to make than and as prepositions should be resisted—though no ambiguity can actually result here.

It must further be noticed that both as and than are conjunctions of the sort that can either, like and, &c., merely join coordinates, or, like when, &c., attach a subordinate clause to what it depends on. This double power sometimes affects case.

This example is defensible, as being here a subordinating conjunction, and as he being equivalent to as he is. But it is distinctly felt to need defence, which as him would not; as would be a coordinating conjunction, and simply join the pronoun him to the noun men. So, with than:

On the other hand, we could not say indifferently, I am as good as he, and I am as good as him; the latter would imply that as was a preposition, which it is not. And it is not always possible to choose between the coordinating and the subordinating use. In the next example only the coordinating will do, no verb being capable of standing after he; but the author has not observed this.

A difficult question, however, arises with relatives after than. In the next two examples whom is as manifestly wrong as who is manifestly intolerable:

The only correct solution is to recast the sentences. For instance, ... whose acquaintance with ... is unrivalled among Englishmen; and ... unsurpassed in his day as a preacher. But perhaps the convenience of than whom is so great that to rule it out amounts to saying that man is made for grammar and not grammar for man.

7. Compound possessives.

This is strictly the proper place for drawing attention to a question that has some importance because it bears on the very common construction discussed at some length in the gerund section. This is the question whether, and to what extent, compound possessives may be recognized. Some people say some one else’s, others say some one’s else. Our own opinion is that the latter is uncalled for and pedantic. Of the three alternatives, Smith the baker’s wife, Smith’s wife the baker, the wife of Smith the baker, the last is unmitigated Ollendorff, the second thrusts its ambiguity upon us and provokes an involuntary smile, and the first alone is felt to be natural. It must be confessed, however, that it is generally avoided in print, while the form that we have ventured to call pedantic is not uncommon. In the first of the examples that follow, we should be inclined to change to Nanny the maid-of-all-work’s, and in the second to the day of Frea, goddess of, &c.

Very little comment will be needed; we have only to convince readers that mistakes are common, and caution therefore necessary.

1. The copula should always agree with the subject, not with the complement. These are wrong:

It must be remembered that in questions the subject often comes after the verb and the complement before it; but the same rule must be kept. E. g., if the last example were put as a question instead of as a negative statement, ‘What proof is the inconsistencies?’ would be wrong, and ‘What proof are &c.?’ right.

Some sentences in which the subject contains only, a superlative, &c., have the peculiarity that subject and complement may almost be considered to have changed places; and this defence would probably be put in for the next three examples; but, whether actually wrong or not, they are unpleasant. The noun that stands before the verb should be regarded as the subject, and the verb be adapted to it.

The next example is a curious problem; the subject to were is in sense plural, but in grammar singular (finding, verbal noun):

2. Mistakes in the number of verbs are extremely common when a singular noun intervenes between a plural subject (or a plural noun between a singular subject) and its verb. It is worth while to illustrate the point abundantly; for it appears that real doubt can exist on the subject:—‘“No one but schoolmasters and schoolboys knows” is exceedingly poor English, if it is not absolutely bad grammar’ (from a review of this book, 1st ed.).

This sentence, strictly taken as it stands, would mean something that the writer by no means intends it to, viz., ‘Though the ceremonies are confusing, there is a great deal in them’.

3. They, them, their, theirs, are often used in referring back to singular pronominals (as each, one, anybody, everybody), or to singular nouns or phrases (as a parent, neither Jack nor Jill), of which the doubtful or double gender causes awkwardness. It is a real deficiency in English that we have no pronoun, like the French soi, son, to stand for him-or-her, his-or-her (for he-or-she French is no better off than English). Our view, though we admit it to be disputable, is clear—that they, their, &c., should never be resorted to, as in the examples presently to be given they are. With a view to avoiding them, it should be observed that (a) the possessive of one (indefinite pronoun) is one’s, and that of one (numeral pronoun) is either his, or her, or its (One does not forget one’s own name: I saw one of them drop his cigar, her muff, or its leaves); (b) he, his, him, may generally be allowed to stand for the common gender; the particular aversion shown to them by Miss Ferrier in the examples may be referred to her sex; and, ungallant as it may seem, we shall probably persist in refusing women their due here as stubbornly as Englishmen continue to offend the Scots by saying England instead of Britain. (c) Sentences may however easily be constructed (Neither John nor Mary knew his own mind) in which his is undeniably awkward. The solution is then what we so often recommend, to do a little exercise in paraphrase (John and Mary were alike irresolute, for instance). (d) Where legal precision is really necessary, he or she may be written in full. Corrections according to these rules will be appended in brackets to the examples.

4. Other mistakes involving number made with such pronominals, or with nouns collective, personified, or abstract.

When each is appended in apposition to a plural subject, it should stand after the verb, or auxiliary, which should be plural; read here, contain each, wish each in turn (or, each of whom wishes in turn).

Every one is indeed a host in himself, if he is the leading maritime nation.

The personification of Japan must be kept up by her.

My notes may be a superfluity; many of my notes may be superfluous, or superfluities; or many a note of mine may be a superfluity; but it will hardly pass as it is.

5. Though nouns of multitude may be freely used with either a singular or a plural verb, or be referred to by pronouns of singular or plural meaning, they should not have both (except for special reasons and upon deliberation) in the same sentence; and words that will rank in one context as nouns of multitude may be very awkward if so used in another.

6. Neither, either, as pronouns, should always take a singular verb—a much neglected rule. So also every.

Were, however, may be meant for the subjunctive, when it would be a fault of style, not of grammar.

7. For rhetorical reasons, a verb often precedes its subject; but enthusiasm, even if appropriate, should not be allowed to override the concords.

The chief point that requires mention is ill treatment of the more. In this phrase the is not the article, but an adverb, either relative or demonstrative. In the more the merrier it is first relative and then demonstrative: by-how-much we are more, by-so-much we shall be merrier. When the relative the is used, it should always be answered regularly by, or itself answer, the demonstrative the. Attempts to vary the formula are generally unhappy; for instance,

This should run, perhaps: whose efficiency impressed him the more, the more he knew him—though it must be confessed that the double form is nearly always uncomfortable if it has not the elbowroom of a whole sentence to itself. That, however, is rather a question of style than of syntax; and other examples will accordingly be found in the section of the Chapter Airs and Graces concerned with originality.

Most readers will feel that this is an uncomfortable compromise between The farther we advance the more do we see and As we advance we see confusion more and more unfold itself. Similarly,

But it is when the demonstrative is used alone with no corresponding relative clause—a use in itself quite legitimate—that real blunders occur. It seems sometimes to be thought that the more is merely a more imposing form of more, and is therefore better suited for a dignified or ambitious style; but it has in fact a perfectly definite meaning, or rather two; and there need never be any doubt whether more or the more is right. One of the meanings is a slight extension of the other. (1) The correlative meaning by so much may be kept, though the relative clause, instead of formally corresponding and containing the (meaning by how much) and a comparative, takes some possibly quite different shape. But it must still be clear from the context what the relative clause might be. Thus, ‘We shall be a huge crowd’.—‘Well, we shall be the merrier’. Or, ‘If he raises his demands, I grant them the more willingly’, i. e., The more he asks, the more willingly I give. This instance leads to the other possible meaning, which is wider. (2) The original meaning of the demonstrative the is simply by that; this in the complete double form, and often elsewhere, has the interpretation, limited to quantity, of by so much, or in that proportion; but it may also mean on that account, when the relative clause is not present. Again, however, the context must answer plainly in some form the question On what account? Thus, He has done me many good turns; but I do not like him any the better; i. e., any better on that account; i. e., on account of the good turns.

The function of the, then, is to tell us that there is, just before or after, an answer to one of the questions, More by what amount? More on what account? If there is no such answer, we may be sure that the comparative has no right to its the. We start with a sentence that is entitled to its the, but otherwise unidiomatic.

The less is less on that account, viz., that we are reminded. But the preposition required when the cause is given in this construction by a noun is for, not by. Read for the reminder. The type is shown in None the better for seeing you. Our sentence is in fact a mixture between Our depression is not lessened by the reminder, and We are not the less depressed for the reminder; and the confusion is the worse that depressed by happens to be a common phrase.

The tells us that we can by looking about us find an answer either to Not less true by what amount? or to Not less true on what account? There is no answer to the first except Not less true about the friends in proportion as it was truer about Mr. Sowerby; and none to the second except Not less true about the friends because it was true about Mr. Sowerby. Both are meaningless, and the the is superfluous and wrong.

This is such an odd tangle of the two formulae as ... so, the more ... the more, that the reader is tempted to cut the knot and imagine what is hardly possible, that the is meant for the ordinary article, agreeing with kind of criticism understood between the and more. Otherwise it must be cured either by omitting the, or by writing The more valuable his criticism, the more rarely is it met with. If the latter is done, than that of other men will have to go. Which suggests the further observation that the with a comparative is almost always wrong when a than-clause is appended. This is because in the full double clause there is necessarily not a fixed standard of comparison, but a sliding scale. The following example, not complicated by any the, will make the point clear:

You can be more averse than ever, or more and more averse, but not more and more averse than ever. Ever can only mean the single point of time in the past, whichever it was, at which you were most averse. But to be more and more averse is to be more averse at each stage than at each previous stage. Just such a sliding scale is essential with the more ... the more. And perhaps it becomes so closely associated with the phrase that the expression of a fixed standard of comparison, such as is inevitably set up by a than-clause, is felt to be impossible even when the demonstrative the stands alone. In the next two examples, answers to the question More on what account? can be found, though they are so far disguised that the sentences would be uncomfortable, even if what makes them impossible were absent. That is the addition of the than-clause in each.

The the might pass if than Australia were omitted, and there would be no objection to it if we read further (for in the case) if we take the case, and better still, placed that clause first in the sentence: Nor, if we take the case of Canada, is the way any the more open. The then means on that account, viz., because we have substituted Canada.

In this, the answer to More on what account? is possibly implied in the last clause; it would perhaps be, if clearly put, Because I go to church seldomer. The right form would be, I can the more acknowledge ... for going (or that I go) to church only once where twenty years ago I went ten times. Unless the than-clause is got rid of, we ought to have more without the.

This question of the is important for lucidity, is rather difficult, and has therefore had to be treated at length. The other points that call for mention are quite simple; they are illogicalities licensed by custom, but perhaps better avoided. Avoidance, however, that proclaims itself is not desirable; to set readers asking ‘Who are you, pray, that the things everybody says are not good enough for you?’ is bad policy; ‘in vitium ducit culpae fuga si caret arte.’ But if a way round presents itself that does not at once suggest an assumption of superiority, so much the better.

1. More than I can help.

These really mean, of course, more than he (we) cannot help. To say that, however, is by this time impossible. More than he need, if (when) he can help it, too much, unnecessarily, and other substitutes, will sometimes do.

2. Most of any (singular).

Though this is extremely common, as the examples are enough to show, there is seldom any objection to saying either most of all or more than any.

3. Most with words that do not admit of degrees.

Unique has been separately dealt with in the chapter on Vocabulary. Ideal is another word of the same sort; an ideal solution is one that could not possibly be improved upon, and most is nonsense with it; an ideal and most obvious should be read in the example:

a. Defining and non-defining relative clauses.

For the purposes of b. and c. below, all relative clauses are divided into defining and non-defining. The exact sense in which we use these terms is illustrated by the following groups, of which (i) contains defining clauses, (ii) non-defining.

The function of a defining relative clause is to limit the application of the antecedent; where that is already precise, a defining clause is not wanted. The limitation can be effected in more than one way, according to the nature of the antecedent. As a rule, the antecedent gives us a class to select from, the defining clause enables us to make the selection. Thus in our first example the antecedent leaves us to select from the general class of ‘men’, the defining clause fixes the particular man (presumably the only man, or the only man that would occur in the connexion) ‘who called yesterday’. Sometimes, however, the functions of the two are reversed. When we have an antecedent with a superlative, or other word of exclusive or comprehensive meaning, such as ‘all’, ‘only’, ‘any’, we know already how to make our selection, and only wait for the relative clause to tell us from what class to make it. We know that we are to choose ‘the best novel’: the relative clause limits us to the works of Trollope. We are to choose ‘any man’ we like, provided (says our relative clause) that he ‘knows three words of Greek’. In either case, the work of definition is done by the exclusion (implied in the relative clause) of persons or things that the antecedent by itself might be taken to include.

The point to notice is that, whichever way the defining clause does its work, it is essential to and inseparable from its antecedent. If for any reason we wish to get rid of it, we can only do so by embodying its contents in the antecedent: ‘The man in Paris with whom I correspond’ must become ‘My Paris correspondent’. To remove the clause altogether is to leave the antecedent with either no meaning or a wrong one. Even in such extreme cases as ‘the wisest man that ever lived’, ‘the meanest flower that blows’, where the defining clause may seem otiose and therefore detachable, we might claim that future wise men, and past and future flowers, are excluded; but we shall better realize the writer’s intention if we admit that these clauses are only a pretence of limitation designed to exclude the reality; it is as if the writers, invited to set limits to their statements, had referred us respectively to Time and Space.

This fact, that the removal of a defining clause destroys the meaning of the antecedent, supplies an infallible test for distinguishing between the defining and the non-defining clause: the latter can always, the former never, be detached without disturbing the truth of the main predication. A non-defining clause gives independent comment, description, explanation, anything but limitation of the antecedent; it can always be rewritten either as a parenthesis or as a separate sentence, and this is true, however essential the clause may be to the point of the main statement. ‘Jones’, in our last example above, is quoted chiefly as one ‘who should know something of the matter’; but this need not prevent us from writing: ‘Jones thinks differently; and he should know something of the matter’.

To find, then, whether a clause defines or does not define, remove it, and see whether the statement of which it formed a part is unaltered: if not, the clause defines. This test can be applied without difficulty to all the examples given above. It is true that we sometimes get ambiguous cases: after removing the relative clause, we cannot always say whether the sense has been altered or not. That means, however, not that our test has failed, but that the clause is actually capable of performing either function, and that the main sentence can bear two distinct meanings, between which even context may not enable us to decide. The point is illustrated, in different degrees, by the following examples:

The former of these is quite ambiguous. The bringing forward of an amendment (no matter what or whose) may be all that the writer meant to tell us of in the first instance; the relative clauses are then non-defining clauses of description. On the other hand, both clauses may quite well be meant to define; and it is even possible that the second is meant to define, and the first not, though the coordination is then of a kind that we shall show under c. to be improper. Similarly, in the second sentence, ‘to portend developments’ may possibly be complete in itself; the whole might then be paraphrased thus: ‘It was thought that the matter would not stop there: but it has’. More probably the clause is meant to define: ‘It was held to portend what have since proved to be unrealized developments’. This view is confirmed, as we shall see, both by the use of ‘that’ (not ‘which’) and by the absence of a comma before it.

Punctuation is a test that would not always be applicable even if all writers could be assumed to punctuate correctly; but it is often a guide to the writer’s intention. For (1) a non-defining clause should always be separated from the antecedent by a stop; (2) a defining clause should never be so separated unless it is either preceded by a parenthesis indicated by stops, or coordinated with a former defining clause or with adjectives belonging to the antecedent; as in the following examples:

It will occur to the reader that our last two examples are strictly speaking exceptions to the rule of defining clauses, since they tell us only what is already implied, and could therefore be removed without impairing the sense. That is true to some extent of many parallel defining clauses: they are admissible, however, if, without actually giving any limitation themselves, they make more clear a limitation already given or implied; if, in fact, they are offered as alternative versions or as reminders. Our next example is of a defining clause of the same kind:

The limitation given by ‘this’ is repeated in another form by the relative clause. ‘This estimate, the one he gives, is....’

The reader should bear in mind that the distinction between the two kinds of relative is based entirely on the closeness of their relation to the antecedent. The information given by a defining clause must be taken at once, with the antecedent, or both are useless: that given by a non-defining clause will keep indefinitely, the clause being complete in sense without the antecedent, and the antecedent without the clause. This is the only safe test. To ask, for instance, whether the clause conveys comment, explanation, or the like, is not a sufficient test unless the question is rightly understood; for, although we have said that a non-defining clause conveys comment and the like, as opposed to definition of the antecedent, it does not follow that a defining clause may not (while defining its own antecedent) contribute towards comment; on the contrary, it is often open to a writer to throw his comment into such a form as will include a defining clause. It may even appear from a comparison of the two sentences below that this is the origin of the non-defining clause, (2) being an abbreviation of (1):

In (1), a comment is introduced by ‘a man’ in apposition with Lewis; ‘a man’ is antecedent to a defining relative clause; separate them, and the antecedent is meaningless. But next remove the connecting words ‘a man’, and the relative changes at once its antecedent and its nature: the antecedent is ‘Lewis’; the relative is non-defining; and the clause is a comment, and does not merely contribute to one.

b. ‘That’ and ‘who’ or ‘which’.

‘That’ is evidently regarded by many writers as nothing more than an ornamental variation for ‘who’ and ‘which’, to be used, not indeed immoderately, but quite without discrimination. The opinion is excusable; it is not easy to draw any distinction that is at all consistently supported by usage. There was formerly a tendency to use ‘that’ for everything: the tendency now is to use ‘who’ and ‘which’ for everything. ‘That’, from disuse, has begun to acquire an archaic flavour, which with some authors is a recommendation. De Quincey, for one, must certainly have held that in exalted prose ‘that’, in all connexions, was the more dignified relative; his higher flights abound in curious uses of the word, some instances of which are quoted below.

This confusion is to be regretted; for although no distinction can be authoritatively drawn between the two relatives, an obvious one presents itself. The few limitations on ‘that’ and ‘who’ about which every one is agreed all point to ‘that’ as the defining relative, ‘who’ or ‘which’ as the non-defining. We cannot say ‘My father, that left Berlin last night, will shortly arrive’, and an examination of instances would show that we can never use ‘that’ where the clause is unmistakably non-defining. On the other hand, we cannot say ‘All which I can do is useless’; this time, it is true, the generalization will not hold; ‘which’ can, and sometimes must, be used, and ‘who’ commonly is used, in defining clauses. But that is explained partly by the obvious inconvenience sometimes attending the use of ‘that’, and partly by the general tendency to exclude it from regular use, which has already resulted in making it seem archaic when used of persons, except in certain formulae.

The rules given below are a modification of this principle, that ‘that’ is the defining, ‘who’ or ‘which’ the non-defining relative; the reason for each modification is given in its place. We must here remind the reader of the distinction drawn in a. between defining and non-defining clauses: a defining clause limits the application of the antecedent, enabling us to select from the whole class to which the antecedent is applicable the particular individual or individuals meant.

1. ‘That’ should never be used to introduce a non-defining clause; it is therefore improperly used in all the following examples:

‘That I represent’ is right: ‘that daily share’ is wrong.

‘That’ and ‘which’ should change places.

If this is a defining clause, excluding ‘the social and economic sciences that’ do not interest the writer, the comma after ‘sciences’ should be removed.

2. ‘Who’ or ‘which’ should not be used in defining clauses except when custom, euphony, or convenience is decidedly against the use of ‘that’. The principal exceptions will be noted below; but we shall first give instances in which ‘that’ is rightly used, and others in which it might have been used with advantage.

‘That’ should have been used in both clauses.

In the following examples, ‘that’ is to be preferred to ‘which’; especially with antecedent ‘it’, and after a superlative or other word of exclusive or comprehensive meaning, such as ‘all’, ‘only’, ‘any’.

‘Which consisted’ is right: but we should have ‘that was held’; the clause defines.

‘Of which’ in a defining clause is one of the recognized exceptions; but we ought to have ‘that subserve’.

3. We come now to the exceptions. The reader will have noticed that of all the instances given in (2) there is only one—the last—in which we recommend the substitution of ‘that’ for ‘who’; in all the others, it is a question between ‘that’ and ‘which’. ‘That’, used of persons, has in fact come to look archaic: the only cases in which it is now to be preferred to ‘who’ are those mentioned above as particularly requiring ‘that’ instead of ‘which’; those, namely, in which the antecedent is ‘it’, or has attached to it a superlative or other word of exclusive meaning. We should not, therefore, in the Spectator instance above, substitute ‘the person that desires’ for ‘who desires’; but we should say

Outside these special types, ‘that’ used of persons is apt to sound archaic.

4. It will also have been noticed that all the relatives in (2) were either in the subjective case, or in the objective without a preposition. ‘That’ has no possessive case, and cannot take a preposition before it. Accordingly ‘the man that I found the hat of’ will of course give place to ‘the man whose hat I found’; and ‘the house in which this happened’ will generally be preferred to ‘the house that this happened in’. The latter tendency is modified in the spoken language by the convenient omission of ‘that’; for always in a defining clause, though never in a non-defining, a relative in the objective case, with or without a preposition, can be dropped. But few writers like, as a general rule, either to drop their relatives or to put prepositions at the end. ‘The friends I was travelling with’, ‘the book I got it from’, ‘the place I found it in’, will therefore usually appear as

5. Euphony demands that ‘that that’ should become ‘that which’, even when the words are separated; and many writers, from a feeling that ‘which’ is the natural correlative of the demonstrative ‘that’, prefer the plural ‘those which’; but the first example quoted in (2) seems to show that ‘those ... that’ can be quite unobjectionable.

6. A certain awkwardness seems to attend the use of ‘that’ when the relative is widely separated from its antecedent. When, for instance, two relative clauses are coordinate, some writers use ‘that’ in the first, ‘which’ in the second clause, though both define. This point will be illustrated in c., where we shall notice that inconsistency in this respect sometimes obscures the sense.

It may seem to the reader that a rule with so many exceptions to it is not worth observing. We would remind him (i) that it is based upon those palpable misuses of the relatives about which every one is agreed; (ii) that of the exceptions the first and last result from, and might disappear with, the encroachment of ‘who’ and the general vagueness about the relatives; while the other two, being obvious and clearly defined, do not interfere with the remaining uses of ‘that’; (iii) that if we are to be at the expense of maintaining two different relatives, we may as well give each of them definite work to do.

In the following subsections we shall not often allude to the distinction here laid down. The reader will find that our rules are quite as often violated as observed; and may perhaps conclude that if the vital difference between a defining and a non-defining clause were consistently marked, wherever it is possible, by a discriminating use of ‘that’ and ‘which’, false coordination and other mishandlings of the relatives would be less common than they are.

c. ‘And who’; ‘and which’.

The various possibilities of relative coordination, right and wrong, may be thus stated: (i) a relative clause may be rightly or wrongly coordinated with another relative clause; this we shall call ‘open’ coordination; (ii) it may be rightly or wrongly coordinated with words that are equivalent to a relative clause, and for which a relative clause can be substituted; ‘latent’ coordination; (iii) a clause that has obviously no coordinate, open or latent, may yet be introduced by ‘and’ or other word implying coordination; for such offenders, which cannot be coordinate and will not be subordinate, ‘insubordination’ is not too harsh a term.

The following are ordinary types of the three classes:

A coordination in which ‘and’ is the natural conjunction may also be indicated simply by a comma; there is safety in this course, since the clause following the comma may be either coordinate or subordinate. But we have to deal only with clauses that are committed to coordination.

‘Insubordination’ will not detain us long; it is always due either to negligence or to gross ignorance; we shall illustrate it in its place with a few examples, but shall not discuss it. With regard, however, to open and latent coordination opinions differ; there is an optimist view of open coordination, and a pessimist view of latent, both of which seem to us incorrect. It is held by some that open coordination (provided that the relatives have the same antecedent) is never wrong, and by some—not necessarily others—that latent coordination is never right: we shall endeavour to show that the former is often wrong, and the latter, however ungainly, often right.

The essential to coordination is that the coordinates should be performing the same function in the sentence. It is not necessary, nor is it enough, that they should be in the same grammatical form: things of the same form may have different functions, and things of different forms may have the same function. If we say ‘Unambitious men, and who have no experience’, ‘unambitious’ and ‘who have no experience’ are not in the same form, but they have the same function—that of specifying the class of men referred to. Their grammatical forms (vocabulary permitting) are interchangeable: a defining adjective can always take the form of a relative clause, and a defining relative clause can often take the form of an adjective: ‘inexperienced men, and who have no ambition’. ‘Unambitious’ is therefore the true grammatical equivalent of ‘who have no ambition’, and latent coordination between it and a relative clause is admissible.

On the other hand, among things that have the same grammatical form, but different functions, are the defining and the non-defining relative clause. A non-defining clause, we know, can be removed without disturbing the truth of the predication; it has therefore no essential function; it cannot therefore have the same function as a defining clause, whose function we know to be essential. It follows that open coordination is not admissible between a defining and a non-defining clause; and, generally, coordination, whether open or latent, is admissible between two defining or two non-defining coordinates, but not between a defining and a non-defining.

Our object, however, in pointing out what seems to be the true principle of relative coordination is not by any means to encourage the latent variety. It has seldom any advantage over full coordination; it is perhaps more apt to lead to actual blunders; it is usually awkward; and it does violence—needless violence, as often as not—to a very widespread and not unreasonable prejudice. Many writers may be suspected of using it, against their better judgement, merely for the purpose of asserting a right; it is their natural protest against the wholesale condemnation of ignorant critics, who do not see that latent coordination may be nothing worse than clumsy, and that open coordination may be a gross blunder. For the benefit of such critics it seems worth while to examine the correctness of various examples, both open and latent; on the other merits and demerits of the latent variety the reader will form his own judgement.

(i) Open coordination.

All these are correct: in the first both clauses are non-defining, in the others both define.

Wrong: the first clause defines, the second not.

Wrong. The first clause defines, the second is obviously one of comment: the ‘scene’ is not distinguished from those that the leader did take steps to avert.

Wrong. The communes that ‘will not permit’ are not meant to be distinguished from those that will. The second clause is comment, the first defines.

All these are wrong. Thackeray would probably have been saved from these false coordinations if he had observed the distinction between ‘that’ and ‘which’: ‘In the best French (that) he could muster, which in sooth was...’.

Probably the second clause, like the first, is meant to define: if so, the coordination is right; if not, it is wrong. We have alluded to the tendency to avoid ‘that’ when the relative is widely separated from its antecedent; here, the result is ambiguity.

Wrong: ‘what (that which)’ defines, the ‘and which’ clauses do not.

(ii) Latent coordination, between relative clause and equivalent, is seldom correct when the relative clause is non-defining; for the equivalent, with few and undesirable exceptions, is always a defining adjective or phrase, and can be coordinate only with a defining clause. The equivalent must of course be a true one; capable, that is, of being converted into a relative clause without altering the effect of the sentence. Neglect of this restriction often results in false coordination, especially in one particular type of sentence. Suppose that a historian, after describing some national calamity, proceeds: ‘In these distressing circumstances....’ Here we might seem to have two possible equivalents, ‘these’ and ‘distressing’. First let us expand ‘these’ into a relative clause: ‘In the distressing circumstances that I have described’. This, in the context, is a fair equivalent, and as often as not would actually appear instead of ‘these’. But next expand ‘distressing’: ‘In these circumstances, which were distressing’, a non-defining clause. To this expansion no writer would consent; it defeats the object for which ‘distressing’ was placed before the antecedent. That object was to record his own sensibility without disparaging the reader’s by telling him in so many words (as our relative clause does) that the circumstances were distressing; and it is secured by treating ‘distressing’ not as a separate predication but as an inseparable part of the antecedent. ‘Distressing’, it will be observed, cannot give us a defining clause; it is obviously meant to be co-extensive with ‘these’; we are not to select from ‘these’ circumstances those only that are ‘distressing’. Moreover, as ‘these’, although capable of appearing as a relative clause, can scarcely require another relative clause to complete the limitation of the antecedent, it follows that in sentences of this form coordination will generally be wrong. We have examples in the Cowper quotation below, and in the anonymous one that precedes it.

‘Rules that shall be general, and that can....’

‘The man who was goldsmith to ... and who’.

All these are correct, with defining coordinates throughout.

‘Who has ... views, and who....’ ‘Sic’ is the comment of the Times writer. The coordination is correct.

These are the instances of false expansion alluded to above. The former is based on the non-defining expansion ‘in all its moods, which are varied and capricious’; the true expansion being ‘in all the varied and capricious moods in which it reveals itself’, a defining clause, which will not do with the ‘and which’. Similarly, the second is based on the non-defining expansion ‘in my uncle’s present state, which is an infirm one’; the true expansion is ‘in the infirm state in which my uncle now is’. In both, a non-defining clause is coordinated with words that can only yield a defining clause.

If the writer means us to distinguish, among the innovations introduced by the Tudors, those that had also been taken away, the ‘and which’ clause defines, and the coordination is right. But more probably the clause conveys independent information; the coordination is then wrong.

Wrong. ‘A plural substantive’ can yield only the defining clause ‘a substantive that is plural’. Now these words contain an inference from a general grammatical principle (that a plural verb must have a plural subject); and any supplementary defining clause must also be general, not (like the ‘and which’ clause) particular. We might have, for instance, ‘Amabant, being plural, and finite, must have for its subject a plural substantive, and which is in the nominative case’. But the ‘and which’ clause is evidently non-defining; the inference ends at ‘substantive’; then comes the application of it to the particular case.

Wrong. The ‘and which’ clause is non-defining; none of the three possible antecedents (‘Theory’, ‘limit’, ‘imposition’) will give a non-defining clause.

This illustrates an important point. ‘Of Europe’ gives the defining clause ‘that prevails in Europe’; the coordination therefore requires that the ‘and which’ clause should define. Now a defining clause must contain no word that is not meant to contribute to definition; if, then, the ‘and which’ clause defines, the writer wishes to distinguish the religion in question, not only from those European religions that have not been colonially introduced into America, but also from those European religions that have been introduced, but whose introduction is not a matter for regret; that is the only defining meaning that ‘unhappily’ can bear, and unless we accept this interpretation the clause is non-defining.—We shall allude to this sentence again in d., where the possibilities of parenthesis in a defining clause are discussed.

One of our ‘few and undesirable exceptions’, in which the clause-equivalent is non-defining (‘who was already on the spot’); for a person’s name can only require a defining clause to distinguish him from others of the same name. The sentence is an ugly one, even if we remove the ‘and who’ clause; but the coordination is right.

(iii) Insubordination.

To these we may add examples in which the coordinated relatives have different antecedents. In practice, nothing can justify such coordination: in theory, it is admissible when the antecedents are coordinate, as in the following sentence:

But in the following instances, one antecedent is subordinate to another in the same clause, or is in a clause subordinate to that of the other.

It was of Mr. Lovelace that the uncles were afraid.

d. Case of the relative.

Special attention was not drawn, in the section on Case, to the gross error committed in the following examples:

These writers evidently think that in ‘whom we think are hungry’ ‘whom’ is the object of ‘we think’. The relative is in fact the subject of ‘are’; and the object of ‘we know’ is the clause ‘who are hungry’; the order of the words is a necessary result of the fact that a relative subject must stand at the beginning of its clause.

(The same awkward necessity confronts us in clauses with ‘when’, ‘though’, &c., in which the subject is a relative. Such clauses are practically recognized as impossible, though Otway, in a courageous moment, wrote:

Some writers, with a consistency worthy of a better cause, carry the blunder into the passive, renouncing the advantages of an ambiguous ‘which’ in the active; for in the active ‘which’ of course tells no tales.

‘Which it was anticipated would be’. Similarly, the passive of ‘men whom we-know-are-honest’ is the impossible ‘men who are-known-are-honest’: ‘men who we know are honest’ gives the correct passive ‘men who it is known are honest’.

Nor must it be supposed that ‘we know’ is parenthetic. In non-defining clauses (Jones, who we know is honest), we can regard the words as parenthetic if we choose, except when the phrase is negative (Jones, who I cannot think is honest); but in a defining clause they are anything but parenthetic. When we say ‘Choose men who you know are honest’, the words ‘you know’ add a new circumstance of limitation: it is not enough that the men should in fact be honest; you must know them to be honest; honest men of whose honesty you are not certain are excluded by the words ‘you know’. Similarly, in the Guernsey Advertiser quotation above, the writer does not go the length of saying that these are the only gentlemen who would work: he says that they are the only ones of whom he feels sure. The commas of parenthesis ought therefore to go, as well as the comma at ‘island’, which is improper before a defining clause.

The circumstances under which a parenthesis is admissible in a defining clause may here be noticed.

(i) When the clause is too strict in its limitation, it may be modified by a parenthesis:

A whole class, excluded by the defining clause, is made eligible by the parenthesis.

(ii) Similarly, a parenthesis may be added to tell us that within the limits of the defining clause we have perfect freedom of choice:

They must have held office, that is all; it does not matter when.

(iii) Words of comment, indicating the writer’s authority for his limitation, his recognition of the sentiments that it may arouse, and the like, properly stand outside the defining clause: when they are placed within it, they ought to be marked as parenthetic.

The latter sentence is an adaptation of one considered above on p. 91. ‘Unhappily’ there appeared not as a parenthesis but as an inseparable part of the relative clause, which was therefore defining or non-defining, according as ‘unhappily’ could or could not be considered as adding to the limitation. But with the altered punctuation ‘unhappily’ is separable from the relative clause, which may now define: ‘that obtains in Europe and (I am sorry to have to add) in America.’

In sentences of this last type, the parenthesis is inserted in the defining clause only for convenience: in the others, it is an essential, though a negative, part of the definition. But all three types of parenthesis agree in this, that they do not limit the antecedent; they differ completely from the phrases considered above, which do limit the antecedent, and are not parenthetic.

e. Miscellaneous uses and abuses of the relative.

(i) A relative clause is sometimes coordinated with an independent sentence; such coordination is perhaps always awkward, but is not always incorrect. The question arises chiefly when the two have a common subject expressed only in the relative clause; for when the subject is expressed in both, the independent sentence may be taken to be coordinate, not with the relative clause, but with the main sentence to which the relative clause is attached, as in the following instance:

Retain ‘I’, and ‘I was’ may be coordinate with ‘he had left’: remove it, and the coordination is necessarily between ‘I was’ and ‘I felt’. In our next examples the writers are committed:

The first two of these are wrongly coordinated: the third, a curiosity in other respects, is in this respect right. The reason is that in the first two we have a defining, in the third a non-defining relative clause. A defining clause is grammatically equivalent to an adjective (‘violated laws’, ‘the popular idea’), and can be coordinated only with another word or phrase performing the same function; now the phrase ‘we have been receiving’, not being attached to the antecedent by means of a relative, expressed or understood, is not equivalent to an adjective. We could have had ‘and (which we) have been properly punished for neglecting’, or we could have had the ‘and’ sentence in an adverbial form, ‘with the fitting result’; but coordination between the two as they stand is impossible.

The Burke sentence is a worse offender. Coordination of this kind is not often attempted when the antecedent of the relative is subject of the main sentence; and when it is attempted, the two coordinates must of course not be separated by the predicate. If we had had ‘the idea which mankind most commonly conceive of proportion, and very seldom trouble themselves about anything further’, the coordination would have been similar to the other, and could have been rectified in the same way (‘and beyond which they very seldom ...’, or ‘to the exclusion of any other considerations’). But this alteration we cannot make; for there is a further and an essential difference. The Daily Telegraph writer evidently meant his second coordinate to do the work of a defining clause; he has merely failed to make the necessary connexion, which we supply, as above, either by turning the words into a second defining clause, or by embodying them, adverbially, in the first. Burke’s intention is different, and would not be represented by our proposed alteration in the order. All that a defining clause can do in his sentence is to tell us what idea is going to be the subject. If we were to give a brief paraphrase of the whole, italicizing the words that represent the second coordinate, it would be, not ‘mankind’s sole idea of proportion is the suitableness ...’, but ‘mankind’s idea of proportion is the suitableness ..., and very little else’; for the question answered is, not ‘what is mankind’s sole idea?’ but ‘what is mankind’s idea?’ In other words, the second coordinate belongs in intention not, like the relative clause, to the subject, but to the predicate; to rectify it, we must either make it part of the predicate (‘and is not concerned with ...’), or, by inserting ‘they’, coordinate it with the main sentence. Obvious as the latter correction is, the sentence repays close examination, as illustrating the incoherence of thought that may underlie what seems a very trifling grammatical slip.

But in our third example, the relative clause is non-defining; it is grammatically equivalent to, and could be replaced by, an independent sentence: ‘Many persons are in the habit of using it’. There is nothing grammatically wrong in this type of coordination; it is objectionable only because it seems to promise what it does not fulfil. When the common subject of two coordinates is expressed only with the first, it is natural to assume that all words preceding it are also to be applied to both coordinates; and the violation of this principle, though not of course ungrammatical, is often felt to be undesirable in other than relative clauses.

(ii) In the sentences considered above, the antecedent of the relative did not belong to the second coordinate, and could not have been represented in it without the material alterations there proposed. But it may also happen that the antecedent, as in the following examples, belongs equally to both coordinates, being represented in the first by a relative, in the second by some other pronoun.

In the pair of parallel coordinates from Mr. Bryce, insert the suppressed ‘was’, and it becomes clear that ‘whose’, not ‘their’, is the right pronoun.

In the ‘Pizarro’ sentence, ‘it’ is not only superfluous, but disturbing to the reader, who assumes that ‘which’ is common to both clauses, and on reaching ‘it’ has to glance back and check the sentence. Here, as often, the pronoun seems to be added to restore an ill-balanced sentence; but that can be done in several other ways. In the Richardson sentence also the ‘it’ should go.

More commonly, the repetition of the antecedent in another form results from the superstitious avoidance of a preposition at the end:

‘To (which)’ is not common to both coordinates: accordingly the writer finds it necessary to give ‘it’ in the second. But, even if we respect our superstition, and exclude ‘which Sweden will not assent to, but will not go to war to prevent’, we have still the two possibilities of (1) complete relative coordination, ‘to ..., but which ...’; (2) subordination, ‘though she will not go to war to prevent it’.

In our next example, Lord Rosebery, again for fear of a preposition at the end, falls into the trap clumsily avoided by the Times writer:

So perhaps Bagehot, though his verb may be conceive of:

(iii) When the relative is the subject of both coordinates, or the object of both, its repetition in the second is a matter of choice. But to omit the relative when it is in a different case from the first is a gross, though not uncommon, blunder. The following are instances:

It is just conceivable that in the last of these the subject of ‘are’ is ‘we’: if so, the sentence is to be referred to (i) above (wrong coordination of an independent sentence with a defining relative clause).

It is not easy to see why the relative more than other words should be mishandled in this way; few would write (but see p. 61, s. f.) ‘This league we kept and has proved advantageous’.

The condensed antecedent-relative ‘what’ is only an apparent exception to this universal rule. In the sentence ‘What I hold is mine’, ‘what’ is only object to ‘hold’, not subject to ‘is’; the subject to ‘is’ is the whole noun-clause ‘what I hold’. Sentences of this type, so far from being exceptions, often give a double illustration of the rule, and leave a double possibility of error. For just as a single ‘what’ cannot stand in different relations to two coordinate verbs in its clause, so a single noun-clause cannot stand in different relations to two coordinate main verbs. We can say ‘What I have and hold’, where ‘what’ is object to both verbs, and ‘what is mine and has been fairly earned by me’, where it is subject to both; but we cannot say ‘what I have and has been fairly earned by me’. Similarly, we can say ‘What I have is mine and shall remain mine’, where the noun-clause ‘what I have’ is subject to both verbs, and ‘What I have I mean to keep, and will surrender to no man’, where it is object to both; but not ‘What I have is mine, and I will surrender to no man’. Of the various ways of avoiding this error (subordination, adaptation of verbs, insertion of a pronoun, relative or otherwise), that chosen by Miss Brontë below is perhaps the least convenient. Her sentence is, however, correct; that from the Spectator is not.

‘Things that were once realities, and that I long have thought decayed’; a pair of defining clauses.

The condensed ‘what’ must of course be distinguished from the ‘what’ of indirect questions, which is not relative but interrogative. In the following example, confusion of the two leads to an improper coordination.

In the first sentence, ‘what’ is an interrogative, in the second, a condensed antecedent-relative, standing for ‘the sums that’. It is the sums that were enormous: it is the answer to the question ‘What sums did he make?’ that can only be conjectured. The mistake is possible only because ‘can’ and ‘must’ do not reveal their number: ‘can’ is singular, ‘must’ plural.

The differentiation between the two whats and their equivalents is not, indeed, complete: just as the condensed antecedent-relative resembles in form, though not in treatment, the unresolved interrogative, so the interrogative, by resolution into ‘the ... that (which)’, not only resembles, but is grammatically identified with, the uncondensed relative and antecedent. The resolution is, no doubt, convenient: it should be noticed, however, that the verbs with which alone it can be employed (verbs that may denote either perception of a fact or other kinds of perception) are precisely those with which ambiguity may result. ‘I know the house (that) you mean’: it may (antecedent and relative) or may not (resolved interrogative) follow that I have ever seen it. ‘We must first discover the scoundrel who did it’; antecedent and relative? then we must secure the scoundrel’s person; resolved interrogative? then only information is needed. ‘I can give a good guess at the problem that is puzzling you’: and the solution?—I know nothing of the solution; I was resolving an interrogative.

This, however, does not affect sentences like the Macaulay one above: for although the resolved or uncondensed forms (‘the ... which’) are grammatically identified, the condensed or unresolved forms (‘what’) are not.

(iv) The omission of the relative in isolated clauses (as opposed to coordinates) is a question not of correctness but of taste, so far as there is any question at all. A non-defining relative can never be omitted. The omission of a defining relative subject is often effective in verse, but in prose is either an archaism or a provincialism. It may, moreover, result in obscurity, as in the second of our examples, which may possibly puzzle the reader for a moment:

But when the defining relative is object, or has a preposition, there is no limit to the omission, unless euphony is allowed to be one. We give three instances in which the reader may or may not agree that the relative might have been retained with advantage:

(v) When a defining relative has the same preposition as its antecedent, it is not uncommon, in the written as well as in the spoken language, to omit the preposition in the relative clause. There is something to be said for a licence that rids us of such cumbrous formulae as ‘in the way in which’, ‘to the extent to which’, and the like; in writing, however, it should be used with caution if at all.

In the first place, if the preposition is to go, the relative should go too, or if retained should certainly be ‘that’, not ‘which’; and if the verb of the relative clause is the same as in the main sentence, it should be represented by ‘do’, or (in a compound tense) by its auxiliary component.

‘In a way no other book in the world could’: ‘in the way (that) Sir E. Satow has done’.

A further limitation is suggested by our next example:

Here, of course, the relative cannot be omitted, since relative and antecedent are one. But that is not the principal fault, as will appear from a resolution of the antecedent-relative: ‘they are ending with the very thing (that) they ought to have begun ...’. We are now at liberty to omit our relative or retain it, as we please; in either case, the omission of ‘with’ is unbearable. The reason is that ‘with’ does not, like the ‘in’ of our former examples, introduce a purely adverbial phrase: it is an inseparable component of the compound verbs ‘end-with’ and ‘begin-with’, of which the antecedent and relative are respectively the objects. Similarly, we cannot say ‘He has come to the precise conclusion (that) I thought he would come’, because we should be mutilating the verb to ‘come-to’; we can, however, say ‘to the conclusion (that) I thought he would’, ‘come-to’ being then represented by ‘would’.

Finally, the omission is justifiable only when antecedent and relative have the same preposition. Sentences like the next may pass in conversation, but (except with the one noun way) are intolerable in writing:

(vi) The use of ‘such ... who (which)’, ‘such ... that (defining relative)’, for ‘such ... as’ is sometimes an archaism, sometimes a vulgarism.

f. ‘It ... that.’

Two constructions, closely allied, but grammatically distinct, are often confused: (i) Antecedent ‘it’ followed by a defining relative clause with ‘that’ (who, which); (ii) ‘it’ followed by a clause in apposition, introduced by the conjunction ‘that’. The various correct possibilities are represented in the set of examples given below. Relative clauses are marked R, conjunction clauses C. One impossible example is added in brackets, to mark the transition from relative to conjunction.

In the relative construction, the antecedent ‘it’ is invariable, whatever the number and gender of the relative. The main verb is also invariable in number, but in tense is usually adapted to past, though not (for euphony’s sake) to future circumstances: ‘it was you that looked foolish’, but ‘it is you that will look foolish’.

In both constructions, the ‘that’ clause, supplemented or introduced by ‘it’, gives us the subject of a predication, the relative clause (with it) being equivalent to a pure noun, the conjunction clause to a verbal noun in apposition, partly retaining its verbal character. In both, also, the predication answers an imaginary question, recorded distinctly in the relative, less distinctly in the conjunction clause. ‘What do you want?’ ‘It (the thing) that I want is money.’ ‘To whom did you give it?’ ‘It (the persons) that I gave it to was your friends.’ ‘As to your cutting it: give particulars.’ ‘It—that I cut it (my cutting it)—was with a knife.’

From the above examples it will be seen that the two constructions largely overlap. When (as in 1, 2, 5, 8) the relative is subject or direct object of the clause-verb, or is in the possessive case, it cannot be replaced by the conjunction; but when its relation to the clause-verb is marked by a preposition, the conjunction always may take its place, and sometimes must, as in 12 and 13. For the relative clause can only be used when the question reflected in it is calculated to secure the right kind of answer. Now the natural answer to the question ‘What did you cut it with?’ is not ‘difficulty’ but ‘a knife’. The misleading ‘with’ is therefore removed from the relative clause in 13, and placed within the predicate, the definite question ‘What did you cut it with?’ giving place to the vague demand for particulars. ‘With’ being removed, the relative clause falls to pieces, for want of a word to govern the relative, and the conjunction clause takes its place. In the same way, ‘it was a cab (but not high indignation) that he drove away in’; ‘it was a concert (but not curiosity) that I was returning from’; ‘it was a beech-tree (but not unpleasant circumstances) that I found him under’. And, generally, it will be found that a preposition is admissible in the relative clause only when used in the literal or the most obvious sense.

The conjunction clause is, as we have said, a verbal noun; so far a noun that things can be predicated of it, and so far a verb that the things predicated of it are verbal relations and verbal circumstances, indirect object, agent, instrument, means, manner, cause, attendant circumstances; anything but subject and direct object. ‘My giving was to you’; ‘my offering was provisionally’; ‘my concealing it was because I was ashamed’.

The mistakes that constantly occur in careless writers result from hesitation between the two forms where both are possible. The confusion, however, ought not to arise; for always with a relative clause, and never with a conjunction, the complement of the main predicate (the answer to the suppressed question) is a noun or the grammatical equivalent of a noun. ‘A knife’, ‘Jones’, ‘you’, ‘my friend in Chicago’, ‘the man who lives next door’, are the answers that accompany the relative clause: ‘with a knife’, ‘with difficulty’, ‘to you’, ‘occasionally’, ‘because I was ashamed’, are those that accompany the conjunction.

Examples 15 and 16, though quite recognized types, are really artificial perversions. In 15 the true question and answer in the circumstances would be, not, as the sentence falsely implies, ‘When did the plague break out?’ ‘That too happened in this same spring’, but ‘Were there any other notable events in this spring?’ ‘Yes: the plague broke out’. Impressiveness is given to the announcement by the fiction that the reader is wondering when the plague broke out; in fact, he is merely waiting for whatever may turn up in the history of this spring. In 16 we go still further: the implied question, ‘What were your feelings on receiving a (not the) note ...?’ could not possibly be asked; the information that alone could prompt it is only given in the ‘that’ clause.

It has been pointed out in b. that a relative clause with antecedent ‘it’ particularly calls for the relative ‘that’, in preference to ‘which’, and even to ‘who’. Even when the relative is in the possessive case, ‘that’, which has no possessive, is often retained by transferring to the main predicate the noun on which it depends; 8 thus gives place to 9, even at the risk of ambiguity; for the relative clause now supplies us with the question (not ‘whose hat ...?’ but) but ‘what did you borrow?’ leaving us theoretically in doubt whether Jones’s hat is distinguished from his other property, from other people’s hats, or from things in general.

On the other hand, the two blunders that are most frequently made almost invariably have the relative ‘who’ or ‘which’.

‘To me’ implies a conjunction clause: ‘to whom ...’ is a relative clause. ‘It is to me that...’.

‘To you that’, or ‘you to whom’.

Again a common blunder; not, however, a confusion between the two constructions above, but between one of them (the relative) and a third. The sentence explains why every one seems to prefer Shakespeare to Ouida (they are afraid to say that they like Ouida best). ‘What is the explanation of this?’ ‘It is not the plentifulness of taste, but the rarity of courage, that explains it.’ Or, less clumsily, using the construction that Stevenson doubtless intended: ‘It (the inference to be drawn) is not that taste is plentiful, but that courage is rare.’

It is advisable to make a few remarks on the participle and gerund together before taking them separately. As the word gerund is variously used, we first define it. A gerund is the verbal noun identical in form with any participle, simple or compound, that contains the termination -ing. Thus the verb write has the active participles writing, having written, being about to write, about to write, and the passive participles written, having been written, being written, about to be written, being about to be written. Any of these except written, about to write, about to be written, may be a gerund also; but while the participle is an adjective, the gerund is a noun, differing from other nouns in retaining its power (if the active gerund of a transitive verb) of directly governing another noun.

Both these are of great importance for our purpose. The participle itself, even when confusion with the other cannot occur, is much abused; and the slovenly uses of it that were good enough in Burke’s time are now recognized solecisms. Again, the identity between the two forms leads to loose and unaccountable gerund constructions that will probably be swept away, as so many other laxities have been, with the advance of grammatical consciousness. We shall have to deal with both these points at some length.

It is indeed no wonder that the forms in -ing should require close attention. Exactly how many old English terminations -ing is heir to is a question debated by historical grammarians, which we are not competent to answer. But we may point out that writing may now be (1) participle—I was writing; I saw him writing; writing piously, he acts profanely—, (2) gerund or full verbal noun—I object to your writing that—, (3) hybrid between gerund and participle—I do not mind you writing it—, (4) detached verbal noun—Writing is an acquired art—, (5) concrete noun—This writing is illegible. Moreover, the verbal noun writing has the synonym to write, obligatory instead of it in some connexions, better in some, worse in some, and impossible in others; compare, for instance: I do not like the trouble of writing; I shall not take the trouble to write; the trouble of writing is too much for him; it is a trouble to write; writing is a trouble. The grammatical difficulties, that is, are complicated by considerations of idiom.

In these preliminary remarks, however, it is only with the distinction or want of distinction between participle and gerund that we are concerned. The participle is an adjective, and should be in agreement with a noun or pronoun; the gerund is a noun, of which it should be possible to say clearly whether, and why, it is in the subjective, objective, or possessive case, as we can of other nouns. That the distinction is often obscured, partly in consequence of the history of the language, will be clear from one or two facts and examples.

1. The man is building contains what we should all now call, whether it is so or not historically, a participle or verbal adjective: the house is building (older but still living and correct English for the house is being built) contains, as its remarkable difference of meaning prepares us to believe, a gerund or verbal noun, once governed by a now lost preposition.

2. In He stopped, laughing we have a participle; in He stopped laughing, a verbal noun governed directly by the verb; in He burst out laughing, a verbal noun governed by a vanished preposition.

3. Present usage does not bear out the definite modern ideas of the distinction between participle and gerund as respectively adjective and noun. So long as that usage continues, there are various degrees of ambiguity, illustrated by the three following examples. It would be impossible to say, whatever the context, whether the writer of the first intended a gerund or a participle. In the second, a previous sentence would probably have decided the question. In the third, though grammar (again as modified by present usage) leaves the question open, the meaning of the sentence is practically decisive by itself.

In the second of these, if sitting is a participle, the meaning is that the end will be secured by the Commission, which is described by way of identification as the one sitting in Paris. If sitting is gerund, the end will be secured by the wise choice of Paris and not another place for its scene. If Commission’s were written, there could be no doubt the latter was the meaning. With Commission, there is, by present usage, absolutely no means of deciding between the two meanings apart from possible light in the context. In the third, common sense is able to tell us, though grammar gives the question up, that what is interesting is not the other people who have them, but the question whether other people have them.

We shall, in the section on the gerund, take up the decided position that all gerunds ought to be made distinguishable from participles. We are quite aware, however, that in the first place a language does not remodel itself to suit the grammarian’s fancy for neat classification; that secondly the confusion is not merely wanton or ignorant, but the result of natural development; that thirdly the change involves some inconveniences, especially to hurried and careless writers. On the other hand it is certain that the permanent tendency in language is towards the correct and logical, not from it; it is merely hoped that the considerable number of instances here collected may attract the attention of some writers who have not been aware of the question, and perhaps convince them that the distinction is a useful one, that a writer ought to know and let us know whether he is using a participle or a gerund, and that to abandon the gerund when it cannot be distinguished without clumsiness need cause no difficulty to any but the very unskilful in handling words.

The unattached or wrongly attached participle is one of the blunders most common with illiterate or careless writers. But there are degrees of heinousness in the offence; our examples are arranged from 1. to 8. in these degrees, starting with perfect innocence.

1. Participles that have passed into prepositions, conjunctions, or members of adverbial phrases.

Regarded as participles, these are incorrect. It is not you that consider, but I; not he that sees, but we; not men that roughly speak, but the moralist; not years that look, but philosophic historians; not the Bill that assumes, but the newspaper prophet. The development into prepositions, &c., is a natural one, however; the only question about any particular word of the kind is whether the vox populi has yet declared for it; when it has, there is no more to be said; but when it has not, the process should be resisted as long as possible, writers acting as a suspensive House of Lords; an instance will be found in 4.

Three quotations from Burke will show that he, like others of his time, felt himself more at liberty than most good writers would now feel themselves.

Similar constructions may be found on almost every page of Smollett.

2. Participles half justified by attachment to a pronoun implied in my, your, his, their. These are perhaps better avoided.

3. Mere unattached participles for which nothing can be said, except that they are sometimes inoffensive if the word to be supplied is very vague.

4. Participles that may some day become prepositions, &c.

He must be the Bishop; for the immediately preceding Sir, marking the beginning of the letter, shows that no one else has been mentioned; but if we had given the sentence without this indication, no one could possibly have believed that this was so; referring is not yet unparticipled.

5. An unwary writer sometimes attaches a participle to the subject of a previous sentence, assuming that it will be the subject of the new sentence also, and then finds (or rather is not awake enough to find) himself mistaken. This is a trap into which good writers sometimes fall, and so dangerous to bad writers that we shall give many examples. It is important for the tiro to realize that he has not satisfied the elementary requirements of grammar until he has attached the participle to a noun in the same sentence as itself, not in another. He must also remember that, for instance, I went and he came, though often spoken of loosely as a sentence, is in fact as fully two sentences as if each half of it were ten lines long, and the two were parted by a full stop and not connected by a conjunction.

6. A more obvious trap, and consequently less fatal, is a change from the active construction that may have been intended to a passive, without corresponding alterations. If the writers of the next two had used we must admit instead of it must be admitted, a policy that they put forward, instead of a policy put forward, the participles hesitating and believing would have had owners.

7. Participles that seem to belong to a noun, but do not.

Does causing agree with letters? Then the letters annoy the horses. With stopping? Then stopping causes suffering by stopping (thus). With omnibuses? The horses possibly blame those innocents, but we can hardly suppose a human being, even the writer of the sentence, so illogical. The word thus, however, is often considered to have a kind of dispensing power, freeing its participle from all obligations; so:

It is not love that carries out, but the power that denies love, which is not mentioned.

8. Really bad unattached or wrongly attached participles. The reader will generally find no difficulty in seeing what has led to the blunder, and if he will take the trouble to do this, will be less likely to make similar blunders himself.

The last, as the Times reviewer has noticed, will repay analysis in several ways.

9. The absolute construction is not much to be recommended, having generally an alien air in English; but it is sometimes useful. It must be observed, first, that the case used should now invariably be the subjective, though it was otherwise in old English. Secondly, it is very seldom advisable to make an absolute construction and insert a pronoun for the purpose when the participle might simply be attached in ordinary agreement to a noun already to hand. Thirdly, it is very bad to use the construction, but omit to give the participle a noun or pronoun to itself. These three transgressions will be illustrated, in the same order, by the next three examples. But many of the wrong sentences in 5 above may be regarded as absolute constructions with the subject omitted.

There are three questions to be considered: whether a writer ought to let us know that he is using a gerund and not a participle; when a gerund may be used without its subject’s being expressed; when a gerund with preposition is to be preferred to the infinitive.

1. Is the gerund to be made recognizable? And, in the circumstances that make it possible, that is, when its subject is expressed, is this to be done sometimes, or always?

It is done by putting what we call for shortness’ sake the subject of the gerund (i. e., the word me or my in me doing or my doing) in the possessive instead of in the objective or subjective case.

Take the typical sentence: I dislike my best friend(’s) violating my privacy. It cannot be a true account of the matter to say that friend is the object of I dislike, and has a participle violating attached to it. For (a) we can substitute resent, which never takes a personal object, for dislike, without changing the sense. (b) If we substitute a passive construction, also without changing the sense, we find that dislike has quite a different object—privacy.—I dislike my privacy being violated by my friend. (c) Many of us would be willing to adopt the sentiment conveyed who yet would not admit for a moment that they disliked their best friend even when he intruded; they condemn the sin, but not the sinner.

Violating then is not an ordinary participle. It does not follow yet that it is a gerund. It may be an extraordinary participle, fused into one notion with the noun, so that a friend violating means the-violation-by-a-friend. The Latin scholar here at once puts in the idiom of occisus Caesar, which does not generally mean Caesar after he was killed, as it naturally should, but the killing of Caesar, or the fact that Caesar had been killed. The parallel is close (though the use is practically confined to the passive in Latin), and familiar to all who know any Latin at all. But it shows not so much what the English construction is as how educated people have been able to reconcile themselves to an ambiguous and not very reasonable idiom—not very reasonable, that is, after language has thrown off its early limitations, and got over the first difficulty of accomplishing abstract expression of any kind. The sort of fusion assumed is further illustrated for the Latinist, though not so closely, by the Latin accusative and infinitive. This theory then takes violating for a participle fused into one notion with friend. There are two difficulties.

I. The construction in English is, though in the nature of things not as common, yet as easy in the passive as in the active. Now the passive of violating is either violated or being violated. It is quite natural to say, Privacy violated once is no longer inviolable. Why then should it be most unnatural to say, The worst of privacy violated once is that it is no longer inviolable? No one, not purposely seeking the unusual for some reason or other, would omit being before violated in the second. Yet as participles violated and being violated are equally good—not indeed always, but in this context, as the simpler Privacy sentence shows. The only difference between the two participles (except that in brevity, which tells against being violated) is that the longer form can also be the gerund, and the shorter cannot. The almost invariable choice of it is due to the instinctive feeling that what we are using is or ought to be the gerund. A more convincing instance than this mere adaptation of our original example may be added:

The necessity for our infantry taught and practised is absolutely impossible. But why, if being taught is participle, and not gerund?

II. Assuming that the fused-participle theory is satisfactory and recognized, whence comes the general, though not universal impression among those who, without being well versed in grammar, are habitually careful how they speak and write, that constructions like the following are ignorant vulgarisms?—It is no use he (his) doing it; it is no use him (his) doing it; that need not prevent us (our) believing; excuse me (my) interrupting you; a thing (thing’s) existing does not prove that it ought to exist; I was annoyed by Tom (Tom’s) hesitating; the Tsar (Tsar’s) leaving Russia is significant; it failed through the King (King’s) refusing his signature; without us (our) hearing the man, the facts cannot be got at; without the man (man’s) telling us himself, we can never know. With a single exception for one (not both) of the first two, none of these ought to cause a moment’s uneasiness to any one who was consciously or unconsciously in the fused-participle frame of mind; and if they do cause uneasiness it shows that that frame of mind is not effectively present.

The Fused-Participle Theory, having no sufficient answer to these objections, but seeing that the gerund’s case is also weak, naturally tries a counter-attack:—If on the other hand the gerund theory is satisfactory and recognized, how is it conceivable that people should leave out the possessive ’s in the reckless way they do? To which, however, the Gerund makes reply:—I regret that they do leave it out, but at least we can see how they come to; it is the combined result of a mistake and an inconvenience. The mistake is caused by certain types of sentence in which a real, not a fused participle is so used that the noun and its (unfused) participle give a sense hardly distinguishable from a possessive noun and a gerund. Examples are:

There is of course a difference. For instance, in the example about the Czar, as in a previous one about conceiving Matthew Arnold permitting, the participle has a pictorial effect; it invites us to imagine the physical appearance of these two great men under indignity instead of merely thinking of the abstract indignity, as we should have done if Czar’s and Arnold’s had shown that we had a gerund; but the difference is very fine; the possessive sign might be inserted without practical effect in all these four, and in hundreds like them. And unlearned people may be excused for deducing that the subject of the gerund can be used at pleasure without the possessive sign, while the learned comfort themselves with the fused-participle theory. That is the mistake. The inconvenience is this: it is easy enough to use the possessive adjectives (my, &c.), and to add the possessive sign to most names and many single nouns; but the subject of a gerund is often a long phrase, after which the sign is intolerable. So the mistake (that the gerund may have a subject not marked by the possessive) is eagerly applied to obviating the inconvenience (that long gerund subjects must be avoided). And that is why people drop their possessive ’s, and why you, the Fused Participle, flourish, defrauding both me, the Gerund, and the honest participle. Thus answered, the Fused Participle does not continue the argument, but pleads only that there is room for all three forms.

Before giving some examples to help in the decision, we shall summarize our own opinion. (1) It is not a matter to be decided by appeal to historical grammar. All three constructions may have separate legitimate descents, and yet in the interests of clear thought and expression it may be better for one of them to be abandoned. (2) There are two opposite tendencies at present: among careful writers, to avoid the fused participle (this, being negative, can naturally not be illustrated) and to put possessive signs in slightly uncomfortable places by way of compensation; among slovenly writers, to throw off all limits of length for the subject of the fused participle. (3) Long fused-participle phrases are a variety of abstract expression, and as such to be deprecated. Among the resources of civilization is the power of choosing between different ways of saying the same thing; and literary skill is very much a matter of exercising that power; a writer should recognize that if he cannot get round an ugly fused participle there is still much for him to learn. (4) Opportunities for ambiguity are so abundant in English, owing to the number of words whose parsing depends on context, that all aids to precision are valuable; and it is not too much to expect a writer to know and let us know whether he means a participle or a gerund.

a. That the possessive of all pronouns that have the form should be used instead of the objective or subjective is hardly disputed. Correct accordingly:

b. Examples in which the possessive of nouns might be written without a qualm.

c. All the last set involved what were either actual or virtual names of persons; there is more difficulty with abstract nouns, compound subjects, and words of which the possessive is ugly. Those that may perhaps bear the possessive mark will be put first, and alterations suggested for the others.

The real objection to the possessive here is merely the addition to the crowd of sibilants.

Or, if the names are essential, did not in the least mind how soon Edward and I married.

d. A modern construction called the compound possessive was mentioned at the end of the section on Cases. It is sometimes ugly, sometimes inoffensive; that is a matter of degree and of knowing where to draw the line; there is no objection to it in principle. And the application of it will sometimes help out a gerund. The first quotation gives a compound possessive simply; the second, a gerund construction to which it ought to be applicable; the third and fourth, two to which it can be applied; and the last, one to which it cannot.

The only objection to a possessive mark after successor is that the two commas cannot be dispensed with; we must say when ... took for on the occasion of ... taking. Such a premier’s will certainly pass. In the Spectator sentence, we should ourselves allow union’s; opinions will differ. But to put the ’s after poor in the last sentence would be ridiculous; that sentence must be rewritten—insisting that many English guardians of the poor should make—or else poor-law Guardians’ must be used.

e. Sometimes we can get over the difficulty without abandoning the gerund, by some slight change of order.

If itself’s is objected to, omit itself (or shift it to the end), and write nation’s.

f. But many types of sentence remain that will have to be completely changed if the gerund is to be recognizable. It will be admitted about most of our examples that the change is not to be regretted. The subject of the gerund is italicized in each, to emphasize its length.

For account for ... ending write ascertain why ... ended. The sentence is radically bad, because the essential construction seems complete at collision—a false scent. That, which is one of the worst literary sins, is the frequent result of long fused participles. It is quite practically possible here for readers to have supposed that they were going to be told why the fleets met, and not why the meeting ended as it did. In the remaining sentences, we shall say when there is false scent, but leave the reader to examine it.

The compound possessive—Tokio’s—is tempting, but perhaps overbold. Insert whether after depends on, and write is for being.

Omit being the case.

For the first of write that, omit the second of, and omit having. False scent.

For of ... being write that ... will be, if peace’s cannot be endured.

For the absurdity of ... denouncing write how absurd it was for ... to denounce. False scent, though less deceptive.

Without regard (at all events without sufficient regard) to that of....

For in consequence of ... being write because ... has been. Bad false scent again.

Observe the fused participle within fused participle here; and read thus: So far from its being inequitable that the state should relieve, &c.

After these specimens, chosen not as exceptional ones, but merely as not admitting of simple correction by insertion of the possessive mark, the reader will perhaps agree that the long gerund subject—or rather noun phrase of the fused participle—is a monstrosity, the abolition of which would be a relief to him, and good discipline for the writer.

Two sentences are added to show the chaotic state of present practice. Noticing the bold use of the strict gerund in the first, we conclude that the author is a sound gerundite, faithful in spite of all temptations; but a few pages later comes the needless relapse into fused participle.

The following looks like a deliberate avoidance of both constructions by a writer who is undecided between the two. Its being is what should have been written.

And perhaps a shyness of something’s being shown accounts for the next odd arrangement; it is true that entire recasting is what is called for.

2. When must the subject of the gerund (or infinitive) be expressed, and when omitted?

This is not a controversial matter like the last; the principles are quite simple, and will be accepted; but it is necessary to state and illustrate them because they are often forgotten. As the same mistakes are sometimes made with the infinitive, that is to be considered as included.

Roughly, the subject of the gerund (or infinitive) should be expressed if it is different from, and omitted if it is the same as, the subject of the sentence. To omit it when different is positively wrong, and may produce actual ambiguity or worse, though sometimes there is only a slipshod effect; to insert it when the same is generally clumsy.

No one would say ‘I succeeded to his property upon dying’, because, I being the subject of the sentence, my is naturally suggested instead of the necessary his as subject of the gerund; the his must be inserted before dying, even though the nature of the case obviates ambiguity. To take an instance that will show both sides, the following is correct:

Subject of the sentence, I; subject of the gerund, he; they are different; therefore the he must be expressed, in the shape of his. Now rewrite the main sentence as—John Barnaby heard some plain speech from me, Clementina MacTaggart. The sense is the same; but the his before philandering at once becomes superfluous; it is not yet seriously in the way, because we do not know what is the subject of philandering, the name only coming later. Now rewrite it again as—Then John Barnaby heard some plain speech from ... instead of ... The his is now so clumsy as to be almost impossible.

The insertion of superfluous subjects is much less common than the omission of necessary ones; but three examples follow. The first is a rare and precious variety; the second has no apparent justification; for the third it may be said that the unusual his has the same effect as the insertion of the parenthetic words as he actually does after limiting would have had.

In giving the rule summarily, we used the phrase subject of the sentence. That phrase is not to be confined to the subject of the main sentence, but to be referred instead, when necessary, to the subject of the subordinate clause in which the gerund may stand. For instance:

Here by sitting breaks the rule, though the subject of sitting is the same as that of the main verb sit, because the subject of the clause in which sitting comes is not the good, but brokers, &c. The right way to mend this is not to insert their before sitting—which after all is clumsy, though correct—but to make the good the subject of the clause also, by writing as if they thought that by sitting ... they would make the brokers ... see the error.

And sometimes subject of the sentence is to be interpreted still more freely as the word grammatically dominant in the part of the sentence that contains the gerund. For instance:

Here the dominant word is Bible, to which familiarized belongs. So, though she does happen to be the main subject, her must be inserted because the familiarized phrase removes the gerund from the reach of the main subject.

After these explanations we add miscellaneous instances. It will be seen that transgression of the rule, though it seldom makes a sentence ambiguous enough to deceive, easily makes it ambiguous enough to amuse the reader at wrong moments, or gives an impression of amateurish work. Mistakes are mended, sometimes by inserting the subject of the gerund (or infinitive), sometimes by changing the main subject to make it the same as that of the gerund, sometimes by other recasting.

Lucky, reflects the reader, since horses are not good at affixing breechings. Write the drivers can put ... horses ... without having to affix.

Write if I assumed; or else I should induce certain ... persons to talk. It will be noticed that the mistake here, and often, is analogous to the most frequent form of wrongly attached participle (participle, 5); the writer does not observe that he has practically passed from the sphere of the sentence whose subject was the word that he still allows to operate.

Grammar gives his, i. e., the new vicar’s, as subject of following; it is really either my or the parishioners’. Insert my or our, or write After we (I) have followed.

Insert our, or write if connived at.

Who or what sees? Certainly not this, the main subject. Not even temperaments, which have no eyes. Write Persons of sensitive temperament experienced this, &c.

It is not the interests that cannot let it pass, but the countries. Insert for those countries before to let; or write Both Great Britain and the United States were too closely affected in their interests to let....

Write if the law required this....

Write they subjected her to for she underwent.

Write I value her friendship too highly not to....

Grammar suggests that his mouth—or, if indulgent, that he—is not to hear; but the person meant is one. Write one has gagged for has been gagged.

It is France, not Germany, that should not add. Write without France’s adding.

This is a triumph of inconsequence. Write If peace is the object, it should be remembered that ordinary....

It will have occurred to the reader that, while most of the sentences quoted are to be condemned, objection to a few of them might be called pedantic. The fact is that every writer probably breaks the rule often, and escapes notice, other people’s, his own, or both. Different readers, however, will be critical in different degrees; and whoever breaks the rule does so at his own risk; if his offence is noticed, that is hanging evidence against him by itself; if it is not noticed, it is not an offence. Of saying on page 127 Mistakes are mended sometimes by inserting the subject, we plead Guilty if we were caught in the act, but otherwise Not Guilty.

3. Choice between the gerund with preposition and the infinitive.

It was said in the preliminary section on the Participle and Gerund that writing—the verbal noun or gerund—and to write—the infinitive—are in some sense synonyms; but phrases were given showing that it is by no means always indifferent which of the two is used. It is a matter of idiom rather than of grammar; but this seems the most convenient place for drawing attention to it. To give satisfactory rules would require many more examples and much more space than can be afforded. But something will be gained if students are convinced (1) that many of the mistakes made give sentences the appearance of having been written by a foreigner or one who is not at home with the literary language; (2) that the mistakes are nearly always on one side, the infinitive being the form that should only be used with caution; (3) that a slight change in arrangement may require a change from infinitive to gerund or vice versa.

a. When the infinitive or gerund is attached to a noun, defining or answering the question what (hope, &c.) about it, it is almost always better to use the gerund with of; not quite always, however; for instance, an intention to return, usually, and a tendency to think always.

What hope?—That of being understood. Write it so, and treat all the following similarly:

What accounts for these mistakes is the analogy of forms like: Our design was to return; it is a duty to instruct; man has power to interpret (but the power of interpreting); it is my custom to be dressed.

When, however, the noun thus defined is more or less closely fused into a single idea with the verb that governs it, the infinitive becomes legitimate, though seldom necessary.

Had thoughts means was planning; had a weakness means desired; had the satisfaction, was pleased; made as little scruple, scrupled as little.

Again, an interval between the noun defined and the infinitive or gerund makes the former more tolerable.

Or the infinitive is used to avoid a multiplication of of.

If the noun has the indefinite article the infinitive is better sometimes.

A duty to make: the duty of making. Compare power and the power above.

The following is probably an adaptation (not to be commended) of it is necessary for Russia to secure—for Russia to secure being regarded as a fused infinitive like the Latin accusative and infinitive.

b. Though the gerund with of is the usual construction after nouns, they sometimes prefer the gerund with other prepositions also to the infinitive. The gerund with in should be used, for instance, in the following. But euphony operates again in the first.

The humiliation there would be in submitting; or the humiliation it would be to submit.

c. After verbs and adjectives the infinitive is much more common; but no one will use a gerund where an infinitive is required, while many will do the reverse.

This is to avoid aiming at being; compare the avoidance of double of above.

In advising may have been avoided as ambiguous.

In the common phrase addicted to drink, drink is a noun, not a verb.

d. If a deferred subject, anticipated by it, is to be verbal, it must of course be either the infinitive or a gerund without preposition.

It is unfortunate that the idiomatic use, while it comes by nature to southern Englishmen (who will find most of this section superfluous), is so complicated that those who are not to the manner born can hardly acquire it; and for them the section is in danger of being useless. In apology for the length of these remarks it must be said that the short and simple directions often given are worse than useless. The observant reader soon loses faith in them from their constant failure to take him right; and the unobservant is the victim of false security.

Roughly speaking, should follows the same rules as shall, and would as will; in what follows, Sh. may be taken as an abbreviation for shall, should, and should have, and W. for will, would, and would have.

In our usage of the Sh. and W. forms, as seen in principal sentences, there are elements belonging to three systems. The first of these, in which each form retains its full original meaning, and the two are not used to give different persons of the same tense, we shall call the pure system: the other two, both hybrids, will be called, one the coloured-future, the other the plain-future system. In Old English there was no separate future; present and future were one. Shall and will were the presents of two verbs, to which belong also the pasts should and would, the conditionals should and would, and the past conditionals should have and would have. Shall had the meaning of command or obligation, and will of wish. But as commands and wishes are concerned mainly with the future, it was natural that a future tense auxiliary should be developed out of these two verbs. The coloured future results from the application to future time of those forms that were practically useful in the pure system; they consequently retain in the coloured future, with some modifications, the ideas of command and wish proper to the original verbs. The plain future results from the taking of those forms that were practically out of work in the pure system to make what had not before existed, a simple future tense; these have accordingly not retained the ideas of command and wish. Which were the practically useful and which the superfluous forms in the pure system must now be explained.

Thou shall not steal is the type of shall in the pure system. We do not ordinarily issue commands to ourselves; consequently I shall is hardly required; but we often ask for orders, and therefore shall I? is required. The form of the shall present in the pure system is accordingly:

As to the past tense, orders cannot be given, but may be asked about, so that, for instance, What should I do? (i. e., What was I to do?) can be done all through interrogatively.

In the conditionals, both statement and question can be done all through. I can give orders to my imaginary, though not to my actual self. I cannot say (as a command) I shall do it; but I can say, as a conditional command, I should do it.

I shall and we shall are accordingly the superfluous forms of the present shall in the pure system.

Again, with will, I will meaning it is my will, it is obvious that we can generally state this only of ourselves; we do not know the inside of other people’s minds, but we can ask about it. The present runs, then,

The past tense can here be done all through, both positively and interrogatively. For though we cannot tell other people’s present will, we can often infer their past will from their actions. So (I was asked, but) I would not, and Why would I do it? all through. And similarly in the conditionals, I would not (if I could), &c.

The spare forms supplied by the present will, then, are you will, he will, they will; and these, with I shall, we shall, are ready, when the simple future is required, to construct it out of. We can now give

Rule 1. The Pure System

When Sh. and W. retain the full original meanings of command and wish, each of them is used in all three persons, so far as it is required.

The following examples show most of what we inherit directly from the pure system.

It will be noticed that the last four forms are among those that were omitted as not required by the pure system. Will would rarely be required in second and third person statements, but would of course be possible in favourable circumstances, as in describing habitual action, where the will of another may be inferred from past experience. The last of all is a natural extension of the idiom even to things that have no will. All these ‘habitual’ uses are quite different from I will have my way; and though you will have your way is possible, it always has the ‘habitual’ meaning, which I will have my way is usually without.

All the forms in the above list, and others like them, have three peculiarities—that they are not practically futures as distinguished from presents; that they use Sh. for all persons, or W. for all persons, if the idea is appropriate to all persons; and that the ideas are simply, or with very little extension, those of command or obligation and wish.

The coloured-future system is so called because, while the future sense is more distinct, it is still coloured with the speaker’s mood; command and wish receive extensions and include promise, permission, menace, consent, assurance, intention, refusal, offer, &c.; and the forms used are invariably those—from both Sh. and W.—that we called the practically useful ones in the pure system. That is, we have always

And the conditionals, should and would, should have and would have, are used with exactly the same variations. It will be borne in mind, however, that no clear line of division can be drawn between the pure system and the coloured-future system, since the latter is developed naturally (whereas the plain-future system is rather developed artificially) out of the former. And especially the questions of the coloured future are simply those of the pure system without any sort of modification.

Rule 2. The Coloured-Future System

In future and conditional statements that include (without the use of special words for the purpose) an expression of the speaker’s (not necessarily of the subject’s) wish, intention, menace, assurance, consent, refusal, promise, offer, permission, command, &c.—in such sentences the first person has W., the second and third persons Sh.

The only questions possible here are the asking for orders and the requests already disposed of under Rule 1.

Observe that I would like (which is not English) is not justified by this rule, because the speaker’s mood is expressed by like, and does not need double expression; it ought to be I should like, under Rule 3.

Observe also that I sha’n’t, You will go to your room and stay there, are only apparent exceptions, which will be explained under Rule 3.

The archaic literary forms You shall find, A rogue shall often pass for an honest man, though now affected and pretentious, are grammatically defensible. The speaker asks us to take the fact on his personal assurance.

The forms little required in the pure system, and therefore ready to hand for making the new plain future, were I, and we, shall; you, he, and they, will. These accordingly constitute the plain future, and the corresponding forms of the plain conditional are used analogously. Questions follow the same rule, with one very important exception, which will be given a separate rule (4). We now give

Rule 3. The Plain-Future System

In plain statements about the future, and in the principal clause, result, or apodosis, of plain conditional sentences (whether the subordinate clause, condition, or if-clause, is expressed or not), the first person has Sh., the second and third persons W. Questions conform, except those of the second person, for which see Rule 4.

Some apparent exceptions, already anticipated, must here be explained. It may be said that I shall execute your orders being the speaker’s promise, You will go to your room being the speaker’s command, and Sha’n’t (the nursery abbreviation for I shall not do it) being the speaker’s refusal, these are all coloured futures, so that Sh. and W. should be reversed in each. They are such in effect, but they are not in form. In each, the other form would be possible and correct. The first is a promise only so far as the hearer chooses to take as a promise the plain future or impersonal prophecy; but the speaker emphasizes his obedience by implying that of course, since the order has been given, it will be executed; the matter is settled without his unimportant consent. The other two gain force by the opposite assumption that the speaker’s will and the future are absolutely identical, so that what he intends may be confidently stated as a future fact. In the first example the desired submissiveness, in the other two the desired imperiousness, supercilious or passionate, are attained by the same impersonality.

Before giving the rule for second-person questions, we observe that questions generally follow the rule of the class of statement they correspond to. This was shown in the pure system (Rule 1). There are no questions (apart from those already accounted for by the pure system) belonging to the coloured future (Rule 2). In the plain future (Rule 3), first and third person questions are like the plain-future statements. But second-person questions under the plain future invariably use Sh. or W. according as the answer for which the speaker is prepared has Sh. or W. Care is necessary, however, in deciding what that answer is. In Should (would) you like a bathe? should is almost always right, because the answer expected is almost always either Yes, I should, or No, I should not, the question being asked for real information. It is true that Would you like? is very commonly used, like the equally wrong I would like; but it is only correct when the answer is intended to be given by the asker:—No, of course you would not. A clearer illustration of this is the following sentence, which requires Sh. or W. according to circumstances: Will (shall) you, now so fresh and fair, be in a hundred years nothing but mouldering dust?. This might possibly be asked in expectation of an answer from the person apostrophized—Yes, I shall. Much more probably it would be asked in expectation of the answer from the speaker himself to his own question—Alas! yes, you will. And shall ought to be used for the question only in the first case, will in the second case. Similarly, Ah, yes, that is all very well; but will (shall) you be able to do it? Use will if the answer is meant to be No, of course you will not; shall, if the answer expected is Yes, I shall, or No, I shall not.

In practice, Sh. is more commonly required, because questions asked for information are commoner than rhetorical ones. But observe the common Would you believe it?, Answer, No, of course you would not. Should you believe it?, also possible, would indicate real curiosity about the other person’s state of mind, which is hardly ever felt. Would you believe it?, however, might also be accounted for on the ground that the answer would be No, I would not, which would be a coloured-future form, meaning I should never consent to believe.

Rule 4. Second-person Questions

Second-person questions invariably have Sh. or W. by assimilation to the answer expected.

It may be added, since it makes the application of the rule easier, that the second-person questions belonging not to the plain future but to the pure system are also, though not because of assimilation, the same in regard to Sh. and W. as their answers. Thus Will you come? Yes, I will (each on its merits), as well as Shall you be there? Yes, I shall (assimilation). Should you not have known? Yes, I should (each on its merits; should means ought), as well as What should you think? I should think you were right (assimilation). The true form for all second-person questions, then, can be ascertained by deciding what the expected answer is.

This completes what need be said about principal sentences, with the exception of one important usage that might cause perplexity. If some one says to me ‘You would think so yourself if you were in my position’, I may either answer ‘No, I should not’ regularly, or may catch up his word, and retain the W., though the alteration of person requires Sh. Thus—‘Would I, though? No, I wouldn’t’. Accordingly,

Rule 5. Echoes

A speaker repeating and adapting another’s words may neglect to make the alteration from Sh. to W., or from W. to Sh., that an alteration of the person strictly requires.

We have now all the necessary rules for principal sentences, and can put down a few examples of the right usage, noteworthy for various reasons, and some blunders, the latter being illustrated in proportion to their commonness. The number of the rule observed or broken will be added in brackets for reference. The passage from Johnson with which the correct examples begin is instructive.

Right.

Wrong.

We give first many examples of the mistake that is out of all proportion the commonest—using the coloured future when the speaker’s mood is sufficiently given by a separate word. In the second example, for instance, I would ask the favour would be quite right, and would mean I should like to ask. As it stands, it means I should like to like to ask. The same applies to the other instances, which are only multiplied to show how dangerous this particular form is.

Some examples follow that have not this excuse; and the first two deserve comment—the first because it results in serious ambiguity, the second because it is possibly not wrong.

With will, the meaning must be: We won’t call them three or four to one, because that would be exaggeration. But the meaning is intended to be: We will call them that, and it will be no exaggeration. Shall is absolutely necessary, however, to make it bear that interpretation.

It is possible that this is the use of will described as the ‘habitual’ use—he will often stand on his head—under Rule 1. But this is very rare, though admissible, in the first person of the present. We shall rejoice, or simply we rejoice, would be the plain way of saying it.

In the next two, if ‘I think’, and the if-clause, were removed, the shall and will would stand, expressing resolve according to Rule 2. But with those additions it is clear that prophecy or pure future is meant; and shall and will should be will and shall.

We now proceed to Subordinate Clauses, and first to the Substantival. The word ‘reported’ will mean ‘made indirect’ or ‘subordinated substantivally’, not always actually reported.

Reported statement is quite simple when it is of the pure system or the coloured future; the Sh. or W. of the original statement is retained in the reported form, unaffected by any change of person that the reporting involves. Thus: (Pure system) He forgave me (you, or her), though he said I (you, or she) should not have left him in the lurch like that. (Coloured future) You said I (or he) should repent it; either of these is a report of either You shall repent it or He shall repent it. (Coloured future) You said you (or I said I) would apologize; both are reports of I will apologize.

But with the plain-future system there is difficulty and some inconsistency. The change of person sometimes required by reported speech has almost always the effect here of introducing Sh. if I or we appears in the words as reported, and usually the effect of introducing W. if you, he, or they, appears. The following are all the types in which doubt can arise, except that each of these may occur in either number, and in past or present. The form that would be required by analogy (keeping the original Sh. or W.) is given first, and the one generally used instead is added in brackets. Reporting I shall never succeed, we get

Reporting you will (or he will) never succeed, we get

Even those persons who have generally a just confidence in their own correctness about Sh. and W. will allow that they have some doubt about the first pair; and nearly every one will find W. in the second pair, however reasonable and consistent, intolerable.

If the reader will now go through the four sentences again, and substitute for succeed the phrase do it (which may or may not mean succeed), he will see that the orthodox should and shall of the first pair become actually more natural than the commoner would and will; and that even in the second pair will and would are now tolerable. The reason is that with do it there is risk of confusion with the reported forms of I will never do it and you shall never do it, which are not plain futures, but coloured futures meaning something quite different.

Reported questions present the same difficulties. Again those only are doubtful that belong to the plain future. There, for instance, reporting Shall you do it? we can say by the correct analogy I asked him whether he should; and we generally do so if the verb, as here, lends itself to ambiguity: I asked him whether he would do it is liable to be mistaken for the report of Will you do it?—a request. If on the other hand (as in reporting Shall you be there?) there is little risk of misunderstanding, I asked him whether he would is commoner. And again it is only in extreme cases, if even then, that the original W. can be kept when the report introduces I in place of the original question’s you or he. For instance, the original question being How will he be treated?, it may be just possible to say You had made up your mind how I would be treated, because You had made up your mind how I should be treated almost inevitably suggests (assisted by the ambiguity of making up your mind, which may imply either resolve or inference) that the original question was How shall he be treated?

It would be well, perhaps, if writers who take their responsibilities seriously would stretch a point sometimes to keep the more consistent and less ambiguous usage alive; but for practical purposes the rule must run:

Rule 6. Substantival Clauses.

In these (whether ‘reported’ strictly or otherwise subordinated) pure-system or coloured-future forms invariably keep the Sh. or W. of the original statement or question, unaffected by any change of person. Reports of plain-future forms do this also, if there would be serious danger of ambiguity, but almost always have Sh. in the first person, and usually W. in the second and third persons.

As the division of substantival clauses into indirect (or reported or subordinate or oblique) statements, questions, and commands, is familiar, it may be well to explain that in English the reported command strictly so called hardly exists. In what has the force of a reported command it is in fact a statement that is reported. For instance, He said I was to go, though used as the indirect form of Go, is really the indirect of the statement You are to go. He ordered that they should be released (though the actual words were Be they, or Let them be, released) is formed on the coloured-future statement, They shall be released. It is therefore unnecessary to give special rules for reported command. But there are one or two types of apparent indirect command about which, though there is no danger of error, the reader may feel curious.

a. I stipulate that I shall, you shall, he shall, do it. Why shall in all persons? because the original form is: I (you, he) shall do it, I stipulate that, where shall means am to, are to, is to; that is, it is a pure-system form.

b. I beg that you (or he) will do it. He begs that I will do it. Again the original is pure-system: You (or he) will (i. e., you consent to) do it: that is what I beg. I will (i. e., I consent to) do it: that is what he begs.

c. I beg that I (or he) shall not suffer for it. You begged that I should not suffer for it. Observe that b. has will and a. and c. shall, because it is only in b. that the volition of the subject of shall or will is concerned.

d. I wish you would not sneeze. Before subordination this is: You will not sneeze: that is what I wish. W. remains, but will becomes would to give the remoteness always connected with wish, which is seen also, for instance, in I wish I were instead of I wish I be.

Before going on to examples of substantival clauses, we also register, again rather for the curious than for the practical reader, the peculiar but common use of should contained in the following:

In this use should goes through all persons and is equivalent to a gerund with possessive: that a man should be is the same as a man’s being. We can only guess at its origin; our guess is that (1) should is the remote form for shall, as would for will in d. above, substituted in order to give an effect of generality; and (2) the use of shall is the archaic one seen in You shall find, &c. So: a man shall be afraid of his shadow; that a man should be afraid (as a generally observed fact) is strange.

After each of the substantival clauses, of which examples now follow, we shall say whether it is a reported (subordinated) statement, or question, and give what we take to be the original form of the essential words, even when further comment is unnecessary.

Examples of Sh. and W. in Substantival clauses.

Right.

Statement. The original of the first is I shall be; of the second, she will be. In this and the next three the strictly analogical form that we recommended is kept.

Statement. I should not.

Statement. We should. They would is not ‘reported’.

Statement. I should be.

Statement. We will not. The blundering insertion of the question—perhaps due to some hazy notion of ‘putting the question’—may be disregarded.

Question. When will it end?

Statement. You shall pay.

Statement. Under me it shall be. The reporting word not expressed.

Statement. I shall. Again the reporting word absent.

This is not properly speaking reported speech. But the shall is accounted for by a sort of allusion to a supposed prophecy—every question shall one day be laid at rest. In that prophecy, shall would convey that the prophet gave his personal guarantee for it, and would come under Rule 2. This is not to be confused with the use of shall in indefinite clauses that will be noticed later.

Wrong.

Questions. At what step shall we meet? Which of us will reach? The first is accordingly right, the second wrong. The modern writer—who has been at the pains to use the strictly correct should in the first place rather than the now common would—has not seen, as Richardson did in the first of the right examples, that his two clauses are dissimilar.

Statement. Will survive. It is possible, however, that the original was thought of, or rather felt, as Our sympathy shall survive. But as the effect of that is to give the speaker’s personal guarantee for the truth of the thing, it is clearly not a proper statement to make dependent on the doubtful word hope.

Statement. We shall not, or the Government will not. Probably Lord Lansdowne said we, and that accounts for should. But if The Times chooses to represent we by the Government, it must also represent shall by would.

Statement. We shall never.

Statement. We should. Of these two instances it may be thought that the writers would have made the mistake in the original unsubordinated sentence, instead of its arising in the process of subordination; our experience is, however, that many people do in fact go wrong in subordinate clauses who are alive to the danger in simple sentences.

Statement. They shall be incorporated. We have classed this as wrong on the assumption, supported by the word assurance, that the Prime Minister gave a promise, and therefore used the coloured future, and did not state a fact and use the plain future.

Another type of subordinate clause important for Sh. and W. is the conditional protasis or if-clause. It is not necessary, nor with modern writers usual, to mark the future or conditional force of this separately, since it is sufficiently indicated by the apodosis. For instance, If you come I shall be glad; if you came I should be glad; if you had come I should have been glad. But in formal style or with a slight difference of meaning, it is often superfluously done in the protasis too. Sh. is then used for all persons, as, If he should come, you would learn how the matter stands. So:

But to the rule that the protasis takes shall there are three exceptions, real or apparent; W. is found under the following circumstances:

(1.) An original pure-system or coloured-future W. is not changed to Sh. by being used in subordination to if (or unless). It is retained with its full original force instead of some verb like wish or choose. In If we would believe we might move mountains, the meaning is If we chose to believe, different from that of If we believed or should believe. So

If you consented not to be, or did not insist on being.

If he consented to.

(2.) When the if-clause (though a genuine condition) is incorrectly expressed for the sake of brevity and compresses two verbs into one, the W. proper to the retained verb is sometimes necessarily used instead of the Sh. proper to the verb that, though it contains in strict logic the essential protasis, has been crushed out. Thus: If it will be useless I shall prefer not to do it. It is not the uselessness that is the condition of the preference; for the use or uselessness is subsequent to the decision; it is my conviction of the uselessness; so that the full form would be If I shall be (or am in ordinary speech) convinced that it will be useless, I shall prefer, &c. The following example can be defended on this ground, if never again will he standing for if he shall realize that he will never; the feebleness that decides his not wishing is subsequent to it, and can only condition it if taken in the sense of his anticipation of feebleness.

The next is more difficult only because, besides the compression, the if-clause is protasis not to the expressed main sentence, but to another that is suppressed.

Given fully, this would run: I shall wait for fine weather; (at least I should say so) if (I were sure that) that will ever come.

(3.) When an if-clause is not a condition at all, as for instance where it expresses contrast, and is almost equivalent to although, the ordinary plain-future use prevails. Thus: If annihilation will end our joys it will also end our griefs. Contrast with this the real condition, in: If annihilation shall end (or ends) our joys, we shall never regret the loss of them.

Indefinite clauses, relative or other, bearing the same relation to a conditional or future principal sentence that a conditional protasis bears to its apodosis follow the same rules. Thus Whoever compares the two will find is equivalent to If any one compares; When we have won the battle we can decide that question is equivalent to If ever we have won. Accordingly we can if we choose write Whoever shall compare, and When we shall have won; but we cannot write When we will have won, and must only write Whoever will compare if we distinctly mean Whoever chooses to compare. As there is sometimes difficulty in analysing indefinite clauses of this sort, one or two instances had better be considered.

This is clear enough; it is equivalent to if any one should have ... he was....

Equivalent to If ever it at length becomes. Will is therefore wrong; either becomes, or shall become.

This is a complicated example. The shalls will be right if it appears that each shall-clause is equivalent to a conditional protasis. We may show it by starting at the end as with the house that Jack built and constructing the sentence backwards, subordinating by stages, and changing will to shall as the protases come in; it will be allowed that until means to the time when, and that when may be resolved into if ever. Thus we get: a. One will know. b. None will be a member of the legislature unless one shall know. c. It will distil to the time if ever none shall be a member unless one shall know.

Although think what I will is an indefinite relative clause, meaning practically whatever I think, will here is right, the strict sense being whatever I choose to think. Indeed the time of think is probably not, at any rate need not be, future at all; compare Think what I will, I do not tell my thoughts.

We now give

Rule 7. Conditional protasis and Indefinite Clauses

In the protasis or if-clause of conditional sentences Sh. may be used with all persons. Generally neither Sh. nor W. is used. W. is only used (1) when the full meaning of wish is intended; it may then be used with all persons; (2) when the protasis is elliptically expressed; W. may then be necessary with the second and third persons; (3) when the if-clause is not a real conditional protasis; there is then no reason for Sh. with second and third persons. Indefinite classes of similar character follow the same rules.

A few right but exceptional, and some wrong subordinate clauses may now be added.

Examples of Sh. and W. in Subordinate Clauses.

Right.

In these three we have the archaic shall of personal assurance that comes under Rule 2, and its corresponding conditional, appearing in subordinate clauses. There is no objection to it except that, in modern writers, its context must be such as to exonerate it from the charge of affectation.

This use of Sh. after final relatives is seen, if the compound sentence is resolved, to point to an original coloured future: We long for a fresh struggle; a fresh struggle shall restore (that is, we intend it to restore) our glory.

This is the should seem explained under Rule 1 appearing also as subordinate.

Wrong.

It should never be, but often is, forgotten that when the apodosis of a conditional sentence (with or without expressed protasis) is subordinate it is nevertheless still an apodosis, and has still Sh. in the first, W. in the second and third persons.

All these are wrong; in the last two the mistake is perhaps accounted for by the presence of willingly and like. I would not willingly can indeed be defended at the cost of admitting that willingly is mere tautology, and saying that I would not means I should not consent to, according to Rule 2.

It may be worth while to add that the subordinate apodosis still follows the rule even if it is subordinated to if, so that it is part of the protasis of another conditional sentence. The following, which is of course quite correct, seems, but only seems, to break the rules both for protasis and apodosis: If you would be patient for yourself, you should be patient for me. But we have W. with second person in the protasis because would be patient is also apodosis to the implied protasis if occasion should arise; and the should with second person in the apodosis is not a conditional should at all, but a pure-system should, which would be the same with any person; it means simply you ought, or it would be your duty.

We have purposely refrained until now from invoking the subjunctive, because the word is almost meaningless to Englishmen, the thing having so nearly perished. But on this instance it must be remarked that when conjunctions like lest, which could once or still can take a subjunctive (as lest he die), use a compound form instead, they use the Sh. forms for all persons. It is a matter of little importance, since hardly any one would go wrong in such a sentence.

This has its right and its wrong uses. The right are obvious, and can be left alone. Even of the wrong some are serviceable, if not strictly logical. I hoped to have succeeded, for instance, means I hoped to succeed, but I did not succeed, and has the advantage of it in brevity; it is an idiom that it would be a pity to sacrifice on the altar of Reason. So:

But other perfects, while they are still more illogical than these, differ as little in meaning from the present as the deposuisse, dear to the hearts of elegiac writers ancient and modern, differs from deponere. And whereas there is at least metre, and very useful metre, in deposuisse, there is in our corresponding perfect infinitive neither rhyme nor reason. Thus,

To have taken means simply to take; the implication of non-fulfilment that justified the perfects above is here needless, being already given in I should have liked; and the doubled have is ugly in sound. Similar are

The less excusable that Bagehot has started with the correct to be.

Another very common form, still worse, occurs especially after seem and appear, and results from the writer’s being too lazy to decide whether he means He seems to have been, or He seemed to be. The mistake may be in either verb or both.

Sometimes have is even transferred from the verb with which it would make sense to the other with which it makes nonsense.

In the next, the perfect is wanted; for a child that has been flogged cannot be left unflogged—not, that is, in the past; and the future is not meant.

We add, for the reader’s refreshment rather than for practical purposes, an illustration of where careless treatment of have may end:

These, which cost the schoolboy at his Latin and Greek some weary hours, need not detain us long. The reader passes lightly and unconsciously in his own language over mixtures that might have caused him searchings of heart in a dead one.

But there is one corrupt and meaningless form, apparently gaining ground, that calls for protest. When a clause begins with as if, it must be remembered that there is an ellipse. I treat her as tenderly as if she were my daughter would be in full I treat her as tenderly as I should if she were, &c. If this is forgotten, there is danger in some sentences, though not in this one, of using a present indicative in the place where the verb were stands. So:

This may be accounted for, but not justified, as an attempt to express what should be merely implied, our actual possession of authority.

This is absolutely indefensible so far as is is concerned; depends has the same motive as have in the Richardson.

This is a place where as if should not have been used at all. If it is used, the verb should be seemed, not may seem, the full form being as there would be (rumours). Read suggesting that for as if, and seems for may seem.

A mixture between it apparently intends and as if it intended.

May should be might. As if it may not is made to do the work of as if it might not, as of course it may.

The same rule applies to as though.

The use of true subjunctive forms (if he be, though it happen) in conditional sentences is for various reasons not recommended. These forms, with the single exception of were, are perishing so rapidly that an experienced word-actuary[11] puts their expectation of life at one generation. As a matter of style, they should be avoided, being certain to give a pretentious air when handled by any one except the skilful and practised writers who need no advice from us. And as a matter of grammar, the instinct for using subjunctives rightly is dying with the subjunctive, so that even the still surviving were is often used where it is completely wrong. So

Here the last clause means though perhaps it was only planned with an eye to effect (and therefore has no military significance). But if followed by were not necessarily means that it certainly is. The mistake here results in making the clause look as if it were the protasis to It would be advisable, with which it has in fact nothing whatever to do; it is a note on the words military significance. Write was for were.

The sentence describes a meeting with a person who knew hardly any English; he said good morning, though it was nightfall. A single example may be added of the intrusion of were for was in a sentence that is not conditional.

Were, however, is often right and almost necessary: other subjunctives are never necessary, often dangerous, and in most writers unpleasantly formal. The tiro had much better eschew them.

Instances will be found in Part II of verbs constructed with wrong prepositions or conjunctions. Most mistakes of this kind are self-evident; but the verb ‘doubt’, which is constructed with ‘that’ or ‘whether’ according to the circumstances under which the doubt is expressed, requires special notice. The broad distinction is between the positive, ‘I doubt whether (that)’ and the negative, ‘I do not doubt that (whether)’; and the rule, in order to include implied as well as expressed negatives, questions as well as statements, will run thus:

The word used depends upon the writer’s or speaker’s opinion as to the reasonableness of the doubt, no matter in whose mind it is said to exist or not to exist.

1. If there is nothing to show that the writer considers the doubt an unreasonable one, the word is always ‘whether’, which reminds us that there is a suppressed alternative:

To this part of the rule there is no exception.

2. If it is evident that the writer disapproves of the doubt, the words introducing it amount to an affirmation on his part that the thing doubted is undoubtedly true; the alternative is no longer offered; ‘that’ is therefore the word:

This, however, is modified by 3.

3. The ‘vivid’ use of ‘whether’. When the writer’s point is rather the extravagance of the doubt than the truth of the thing doubted, ‘whether’ is often retained:

Here, according to 2., we ought to have ‘that’, since the writer evidently regards the doubt as absurd. But in the first sentence it is necessary for the force of the illustration that the deplorable condition of the doubter’s mind should be vividly portrayed: accordingly, he is represented to us as actually handling the two alternatives. Similarly, in the second, we are invited to picture to ourselves, if we can, a hesitation so ludicrous in the writer’s opinion. We shall illustrate this point further by a couple of sentences in which again the state of mind of the doubter, not the truth of the thing doubted, is clearly the point, but in which ‘that’ has been improperly substituted for the vivid ‘whether’:

The mistake commonly made is to use ‘that’ for ‘whether’ in violation of 1. ‘Whether’ is seldom used in place of ‘that’, and apparent violations of 2. often prove to be legitimate exceptions of the ‘vivid’ kind. Some of our examples may suggest that when the dependent clause is placed before the verb, ‘that’ appears because the writer had not decided what verb of doubt or denial to use. This is probably the true explanation of many incorrect thats, but is not a sufficient defence. It supplies, on the contrary, an additional reason for adhering to ‘whether’: the reader is either actually misled or at any rate kept in needless suspense as to what is going to be said, because the writer did not make up his mind at the right time how to say it. ‘Whether’ at the beginning at once proclaims an open question: after ‘that’ we expect (or ought to expect) ‘I have no reason to doubt’.

In all the following, ‘whether’ should have been used.

‘Whether’ is never suppressed.

In an uninflected language like ours these are ubiquitous, and it is quite impossible to write tolerably without a full knowledge, conscious or unconscious, of their uses. Misuse of them, however, does not often result in what may be called in the fullest sense blunders of syntax, but mostly in offences against idiom. It is often impossible to convince a writer that the preposition he has used is a wrong one, because there is no reason in the nature of things, in logic, or in the principles of universal grammar (whichever way it may be put), why that preposition should not give the desired meaning as clearly as the one that we tell him he should have used. Idioms are special forms of speech that for some reason, often inscrutable, have proved congenial to the instinct of a particular language. To neglect them shows a writer, however good a logician he may be, to be no linguist—condemns him, from that point of view, more clearly than grammatical blunders themselves. But though the subject of prepositions is thus very important, the idioms in which they appear are so multitudinous that it is hopeless to attempt giving more than the scantiest selection; this may at least put writers on their guard. Usages of this sort cannot be acquired from dictionaries and grammars, still less from a treatise like the present, not pretending to be exhaustive; good reading with the idiomatic eye open is essential. We give a few examples of what to avoid.

1. After adjectives and adverbs.

Insensible to, but unconscious of; indifferent to, but oblivious of

The adjectives different and averse, with their adverbs or nouns, differently, difference, aversion, averseness, call for a few words of comment. There is no essential reason whatever why either set should not be as well followed by to as by from. But different to is regarded by many newspaper editors and others in authority as a solecism, and is therefore better avoided by those to whom the approval of such authorities is important. It is undoubtedly gaining ground, and will probably displace different from in no long time; perhaps, however, the conservatism that still prefers from is not yet to be named pedantry. It is at any rate defensive, and not offensive pedantry, different to (though ‘found in writers of all ages’—Oxford Dictionary) being on the whole the aggressor. With averse, on the other hand, though the Oxford Dictionary gives a long roll of good names on each side, the use of from may perhaps be said to strike most readers as a distinct protest against the more natural to, so that from is here the aggressor, and the pedantry, if it is pedantry, is offensive. Our advice is to write different from and averse to. We shall give a few examples, and add to them two sentences in which the incorrect use of from with other words looks like the result of insisting on the slightly artificial use of it after different and averse.

It will be noticed that to is more than usually uncomfortable when it does not come next to different.

From would in this last be clearly better than to; but between the two would be better than either.

Disagree from (for with), and adverse from (for to), seem to have resulted from the superstition against averse and different to.

2. After verbs.

I content you by submitting: I content myself with saying.

The moon consists of green cheese: virtue consists in being good. Consist of gives a material, consist in a definition. Mr. Balfour’s ‘elegant variation’ (see Airs and Graces) is certainly wrong, though nominalists and realists will perhaps differ about which should have been used in both sentences, and no one below the degree of a metaphysician can pretend to decide between them.

We tamper with, but tinker at, the thing that is to be operated on.

His toiling, or him to toil.

On the spur of the moment, but at a moment’s notice. The motive was, no doubt, to avoid repeating at; but such devices are sins if they are detected.

We acquiesce in, not to, though either phrase is awkward enough with a voice; to is probably accounted for again by the desire to avoid repeating in.

3. After nouns.

I find fault with: I find a fault in. Write in or with, as one or the other phrase is meant.

We assume the right of turning, we take the initiative in turning.

Benefits of the benefactor, but to the beneficiary.

4. Superfluous prepositions, whether due to ignorance of idiom, negligence, or mistaken zeal for accuracy.

This is due to confusion with ‘could occupy itself with no question more profitably than with this’.

5. Necessary prepositions omitted.

Write without to will now pass in commercial letters only; elsewhere, we can say ‘I write you a report, a letter’, but neither ‘I will write you’ simply, nor ‘I wrote you that there was danger’. That is, we must only omit the to when you not only is the indirect object, but is unmistakably so at first sight. It may be said that I write you is good old English. So is he was a-doing of it; I guess is good Chaucerian. But in neither case can the appeal to a dead usage—dead in polite society, or in England—justify what is a modern vulgarism.

6. Compound prepositions and conjunctions.

The increasing use of these is much to be regretted. They, and the love for abstract expression with which they are closely allied, are responsible for much of what is flaccid, diffuse, and nerveless, in modern writing. They are generally, no doubt, invented by persons who want to express a more precise shade of meaning than they can find in anything already existing; but they are soon caught up by others who not only do not need the new delicate instrument, but do not understand it. Inasmuch as, for instance, originally expressed that the truth of its clause gave the exact measure of the truth that belonged to the main sentence. So (from the Oxford Dictionary):

But long before Hamilton’s day the word passed, very naturally, into the meaning, for which it need never have been invented, of since or because. Consequently most people who need the original idea have not the courage to use inasmuch as for it, like Sir W. Hamilton, but resort to new combinations with far. Those new combinations, however, as will be shown, fluctuate and are confused with one another. The best thing we can now do with inasmuch as is to get it decently buried; when it means since, since is better; when it means what it once meant, no one understands it. The moral we wish to draw is that these compounds should be left altogether alone except in passages where great precision is wanted. Just as a word like save (except) is ruined for the poet by being used on every page of ordinary prose (which it disfigures in revenge for its own degradation), so inasmuch as is spoilt for the logician.

We shall first illustrate the absurd prevailing abuse of the compound preposition as to. In each of the following sentences, if as to is simply left out, no difference whatever is made in the meaning. It is only familiarity with unnecessary circumlocution that makes such a state of things tolerable to any one with a glimmering of literary discernment. As to flows from the pen now at every possible opportunity, till many writers seem quite unaware that such words as question or doubt can bear the weight of a whether-clause without help from this offensive parasite.

The first as to in this may pass, though plain to is better.

This is the worst of all. The subject of makes (anticipated in the ordinary way by it) is who says a thing; but the construction is obscured by the insertion of as to. We are forced to suppose, wrongly, that it means what brings me here. Worse than the worst, however, at least more aggressively wrong, is an instance that we find while correcting this sheet for the press:

Another objection to the compound prepositions and conjunctions is that they are frequently confused with one another or miswritten. We illustrate from two sets. (a) The word view is common in the forms in view of, with a view to, with the view of. The first expresses external circumstances, existing or likely to occur, that must be taken into account; as, In view of these doubts about the next dividend, we do not recommend.... The other two both express the object aimed at, but must not have the correspondence, a view to, the view of, upset.

Of these the first is correct; but the sentence it comes in is so typical of the compound-prepositional style that no one who reads it will be surprised that its patrons should sometimes get mixed; how should people who write like that keep their ideas clear? The second should have with a view to. Still more should the third, which is ambiguous as well as unidiomatic; the words used ought to mean seeing that her interests are safeguarded already. The fourth and fifth should again have with a view to (or with the view of).

(b) The combinations with far—as far as, so far as, so far that, in so far as, in so far that, of which the last is certainly, and the last but one probably needless—have some distinctions and limitations often neglected. For instance, as far as must not be followed by a mere noun except in the literal sense, as far as London. So far as and so far that are distinguished by good writers in being applied, the first to clauses that contain a doubtful or varying fact, the other to clauses containing an ascertained or positive fact. So far as (and in so far as), that is, means to whatever extent, and so far that means to this extent, namely that.

Correct; but except that would be much better than save in so far as.

Entirely wrong. It is confused with inasmuch as, and since should be written.

Should be so far succeeded that; we are meant to understand that the fleet does keep outside, though it does not go right away as might be wished.

As should be that. It is not doubtful to what extent or whether the Tsar sent. He did send; that is the only point.

Should probably be so far exceptional that. The point is that there is this amount of the exceptional in them, not that their irregularity depends on the doubtful fact of their not being modifications; the word appear ought otherwise to have been parenthetically arranged.

A mixture of was so far indirect that and was indirect in that.

In so far as leaves the popularity and the pulling through doubtful, which they are clearly not meant to be. It should be so far that.

This is a whole sentence, not a fragment, as might be supposed. But as far as (except in the local sense) must have a verb, finite or infinite. Supply goes.

So far as to admit, or in so far as they would admit; not the mixture. And this distinction is perhaps the only justification for the existence of in so far as by the side of so far as; the first is only conjunction, the second can be preposition as well.

CHAPTER III AIRS AND GRACES — Certain Types of Humour — Elegant Variation — Inversion — Archaism — Metaphor — Repetition — Miscellaneous

Certain types of humour—Elegant variation—Inversion—Archaism—Metaphor—Repetition—Miscellaneous.

Some of the more obvious devices of humorous writers, being fatally easy to imitate, tend to outlive their natural term, and to become a part of the injudicious novice’s stock-in-trade. Olfactory organ, once no doubt an agreeable substitute for ‘nose’, has ceased to be legal tender in literature, and is felt to mark a low level in conversation. No amount of classical authority can redeem a phrase that has once reached this stage. The warmest of George Eliot’s admirers, called upon to swallow some tough morsel of polysyllabic humour in a twentieth-century novel, will refuse to be comforted with parallel passages from Adam Bede. Loyalty may smother the ejaculation that ‘George Eliot knew no better’: it is none the less clear to him that we know better now. A few well-worn types are illustrated below.

a. Polysyllabic humour.

b. Playful repetition.

The author is very much at his ease in the last example; the novice who should yawn in our faces with such engaging candour would render himself liable to misinterpretation.

c. The well-worn ‘flood-of-tears-and-sedan-chair’ pleasantry.

d. Other worn-out phrases of humorous tendency.

These two phrases are commonly employed to carry off a humorous description of which the success is doubted. They are equivalents, in light literature, of the encouragement sometimes offered by the story-teller whose joke from Punch has fallen flat: ‘You should have seen the illustration’. Worthy and gallant are similarly used:

It may be doubted whether any resemblance or contrast, however striking, can make it worth a modern writer’s while to call waiters Ganymedes, waitresses Hebes, postmen Mercuries, cabmen Automedons or Jehus. In Scott’s time, possibly, these phrases had still an agreeable novelty: they are now so hackneyed as to have fallen into the hands of writers who are not quite certain who Ganymede and Hebe were. Thus, there are persons who evidently think that it is rather complimentary to one’s host than otherwise to call him an Amphitryon; and others who are fond of using the phrase ‘l’Amphitryon où l’on dîne’ altogether without point, apparently under the impression that ‘où l’on dîne’ is an alternative version for the use of the uninitiated (‘Amphitryon’, that is to say, ‘one’s host’).

The inverted commas seem to implore the reader’s acceptance of this very battered ornament. One could forgive it more easily, if there were the slightest occasion for its appearance here.

The notion that Mr. Kipling, left to himself, is not competent to bring out all the latent possibilities of this phrase is a mistaken one, and argues an imperfect acquaintance with his works.

We include under this head all substitutions of one word for another for the sake of variety, and some miscellaneous examples will be found at the end of the section. But we are chiefly concerned with what may be called pronominal variation, in which the word avoided is either a noun or its obvious pronoun substitute. The use of pronouns is itself a form of variation, designed to avoid ungainly repetition; and we are only going one step further when, instead of either the original noun or the pronoun, we use some new equivalent. ‘Mr. Gladstone’, for instance, having already become ‘he,’ presently appears as ‘that statesman’. Variation of this kind is often necessary in practice; so often, that it should never be admitted except when it is necessary. Many writers of the present day abound in types of variation that are not justified by expediency, and have consequently the air of cheap ornament. It is impossible to lay down hard and fast rules, but two general principles may be suggested: (1) Variation should take place only when there is some awkwardness, such as ambiguity or noticeable monotony, in the word avoided. (2) The substitute should be of a purely pronominal character, a substitute and nothing more; there should be no killing of two birds with one stone. Even when these two requirements are satisfied, the variation is often worse, because more noticeable, than the monotony it is designed to avoid.

The examples in our first group do not offend against (2): how far they offend against (1), and how far they are objectionable on other grounds, we shall consider in detail.

The expert is gratuitous: He would have done quite well.

His mother’s is not only unnecessary, but misleading: there is a difficulty in realizing that her and his mother, so placed, can be meant to refer to the same person.

That dignitary said might have been omitted, with the full stop before it.

The writer has thoroughly puzzled himself. He cannot call Shakespeare Shakespeare, because there is a Shakespeare just before: he cannot call him he, because six other persons in the sentence have claims upon he: and he ought not to call him the dramatist, because Aeschylus and Sophocles were dramatists too. We know, of course, which dramatist is meant, just as we should have known which he was meant; but the appropriation is awkward in either case. The dramatist is no doubt the best thing under the circumstances; but when matters are brought to such a pass that we can neither call a man by his own name, nor use a pronoun, nor identify him by means of his profession, it is time to remodel the sentence.

Elegant variation is the last thing we should expect here. For what is the writer’s principal object? Clearly, to emphasize the idea of reciprocity by the repetition of names, and by their arrangement. Mr. Chamberlain, Mr. Balfour: Mr. Balfour, Mr. Chamberlain. It is easy enough, so far: ‘If Mr. Chamberlain has been injured by the persistent attachment of Mr. Balfour, Mr. Balfour has been equally injured by that of Mr. Chamberlain’. But that is not all that is required: there is to be the graphic touch; arm is to be locked in arm. Now comes the difficulty: in whose arm are we to lock Mr. Chamberlain’s? in ‘his’? in ‘his’? in ‘his own’? in ‘Mr. Balfour’s’? in ‘that of the Prime Minister’? As the locking of arms is perhaps after all only an elegant variation for clinging, remodelling seems again to be the best way out of the difficulty. Perhaps our simplified form above might serve.

‘The horse’.—An unconscious satirist, of tender years but ripe discernment, parsed ‘animal’ in this sentence as a personal pronoun; ‘it replaced the subject of the sentence’. Journalists (it was explained to her) are equipped with many more personal pronouns than ever get into the grammars.

This is, no doubt, a difficult case. The royal pronoun (His Majesty) does not lend itself to repetition: on the other hand, it is felt that hes, if indulged in at all, must be kept a respectful distance apart; hence The King in the third sentence. We can get rid of it by reading ‘... at Clarence Barracks; returning ...’. But of course that solution would not always be possible.

There is no excuse either for the Monarch or for the Emperor Francis Joseph. ‘He’ could scarcely have been misinterpreted even in the latter sentence.

It is natural that Sir Charles Edward Bernard should be introduced to us under his full name; natural, also, that an abbreviation should be chosen for working purposes. But why two abbreviations? If Sir Charles and he are judiciously employed, they will last out to the end of the longest article, without any assistance from Sir Edward.

Among the instances here given, there is scarcely one in which variation might not have been avoided with a little trouble. There are some, indeed, in which it is not gratuitous; and if in these the effect upon the reader were as negative as the writer’s intention, there would be nothing to complain of. But it is not; the artistic concealment of art is invariably wanting. These elephantine shifts distract our attention from the matter in hand; we cannot follow His Majesty’s movements, for wondering what the King will be called next time; will it be plain Edward VII? or will something be done, perhaps, with ‘the Emperor of India’? When the choice lies between monotonous repetition on the one hand and clumsy variation on the other, it may fairly be laid down that of two undesirable alternatives the natural is to be preferred to the artificial.

But variation of this kind is, at the worst, less offensive than that which, in violation of our second principle above, is employed as a medium for the conveyance of sprightly allusion, mild humour or (commonest of all) parenthetic information.

‘Far from it’, in fact.

The parlour was Mr. Jarvie’s.

This is a favourite newspaper type.

The miscellaneous examples given below (except ‘the former of the last two’) are connected with pronominal variation only so far as they illustrate the same principle of false elegance.

The return to marked is humiliating; we would respectfully suggest characterized.

We have not room to record at length, from the Westminster Gazette, the elegant variety of fortune that attended certain pictures, which (within twenty lines) made, fetched, changed hands for, went for, produced, elicited, drew, fell at, accounted for, realized, and were knocked down for, various sums.

Of all the types of inversion used by modern writers, there is perhaps not one that could not be shown to exist in older English. Ordinary modern usage, however, has retained those forms only in which ancient authority combines with practical convenience; and not all of those. To set aside the verdict of time in this respect is to be archaic. Before using inversion, therefore, the novice should ask himself two questions: is there any solid, practical reason (ornamental reasons will not do) for tampering with the normal order of subject and verb? and does the inversion sound natural?

Throughout this section it must be borne in mind that in all questions of right and wrong inversion the final appeal is not to history, but to the reader’s perception: what sounds right to most modern ears is right for modern purposes. When, under balance inversion, we speak of a true and a false principle, we do not mean to imply that the ‘true’ principle was, historically, the origin of this kind of inversion, or that the ‘false’ is a mistaken analogy from it: all that is meant is that if we examine a collection of instances, those that sound natural will prove to be based upon the ‘true’ principle, and those that do not on the ‘false’.

a. Exclamatory inversion.

This may be regarded as an abbreviated form of exclamation, as if the word ‘How’ had dropped out at the beginning, and a note of exclamation at the end. The inverted order, which is normal in the complete exclamation, sounds natural also in the abbreviated form. The requirements for this kind of inversion are these: (1) The intention must be genuinely exclamatory, so that the full form of exclamation could be substituted without extravagance. (2) The word placed first must be that which would bear the chief emphasis in the uninverted form. It should be observed that this is the only kind of inversion in which the emphatic word, as such, stands at the beginning.

Our first three examples satisfy these conditions, and are unobjectionable. The fourth does not: we could not substitute ‘With what difficulty...!’; nor are the first words emphatic; the emphasis is on ‘conceive’. Yet the inversion is inoffensive, being in fact not exclamatory at all, but a licensed extension of negative inversion, which is treated below.

Exclamatory inversion, like everything else that is exclamatory, should of course be used sparingly.

b. Balance inversion.

The following are familiar and legitimate types:

We give the name of ‘balance’ to this kind of inversion because, although the writer, in inverting the sentence, may not be distinctly conscious of rectifying its balance, the fact that it was ill-balanced before is the true cause of inversion. It is a mistake to say that the words placed first in the above examples are so placed for the sake of emphasis; that is a very common impression, and is responsible for many unlawful inversions. It is not emphasis that is given to these words, it is protection; they are placed there to protect them from being virtually annihilated, as they would have been if left at the end. Look at the last of our examples: how can we call the words ‘Among the guests were’ emphatic, or say that they were placed there for emphasis? They are essential words, they show the connexion, nor could the sentence be a sentence without them; but they are as unemphatic as words could well be.—Why, then (it may be asked), are they put at the beginning? is not this an emphatic position? and does not any unusual position give emphasis?—No: it gives not emphasis but prominence, which is another thing.

Put the sentence back into its original form, and we shall see why inversion was desirable. ‘A, B, C, D, E, F ... Z were among the guests.’ Observe how miserably the sentence tails off; it has no balance. By inverting it, we introduce several improvements. First, we give prominence to the unemphatic predicate, and enable it to discharge its humble office, that of a sign-post, indicating the connexion with what has gone before. Secondly, by giving prominence to the predicate, we give balance to the sentence, which before was top-heavy. Thirdly, we give prominence to the subject, by placing it in an unusual position.

Next take the ‘local option’ sentence. Are the words ‘First on our list’ emphatic? Not if the inverter knows his business. How did it run originally? ‘The question of local option stands first on our list.’ These words might be meant to tell us either of two things: what stood first on the list, or where local option stood. If the inversion is right, they are meant to tell us what stood first. If the other had been meant, then ‘First on the list’ would have been emphatic, and the writer would have left it in its place; but as it is not emphatic, and the other words are, the sentence is top-heavy; he therefore inverts it, thus balancing the sentence, and placing the unemphatic words in a prominent position, where they continue to be unemphatic, but are sure to be noticed. In spoken language, the relative importance of the different parts of a sentence can be indicated merely by the inflexion of the voice; but the balance of the sentence is best maintained, even then, by means of inversion.

It is the same with the other examples. If we restore the St. Matthew quotation to the uninverted form, again we have an answer to either of two questions: What is the basis of the law? and What is the importance of these two commandments? Obviously it is meant as an answer to the latter, and therefore the words that convey that answer are the emphatic words; the others are not emphatic, but merely essential to the connexion; the general importance of the ‘two commandments’, as forming the subject-matter of the whole context, does not in the slightest degree affect their relation to the other words in this particular sentence.

It follows from what has been said that true balance inversion is employed not for the sake of impressiveness, but with the purely negative object of avoiding a bad balance. The data required for its justification are (i) An emphatic subject, carrying in itself the point of the sentence, (ii) Unemphatic ‘sign-post’ words, essential to the connexion, standing originally at the end of the sentence, and there felt to be inadequately placed. The results of the inversion must be (iii) That the sign-post stands at the beginning, (iv) That the subject stands absolutely at the end.

When these four conditions are fulfilled, the inversion, far from being objectionable, may tend greatly to vigour and lucidity. It is liable, of course, to be overdone, but there are several ways of avoiding that: sometimes it is possible to place the sign-post at the beginning without inversion; or the uninverted sentence may be reconstructed, so that the subject no longer carries the emphasis; and, as often as not, a sentence of which the accentuation is theoretically doubtful may in practice be left to the reader’s discernment.

One occasional limitation remains to be mentioned, before we proceed to instances. It applies to those sentences only that have a compound verb: if the compound verb cannot be represented simply by its auxiliary component, the inversion may have to be abandoned, on account of the clumsiness of compound verbs in the middle of an inverted sentence, for to carry the other component to the end would be to violate our fourth rule. Take the type sentence ‘To these causes may be attributed ...’, and first let the subject be ‘our disasters’. The clumsiness of the verb is then distinctly felt; and ‘To these causes may our disasters be attributed’ is ugly enough to show the importance of the rule it violates. But next let the subject be ‘every one of the disasters that have come upon us’. This time the inversion is satisfactory; whence we conclude that if the verb is compound, the subject must be long as well as emphatic, or the inversion will not do.

These are both formed on the right principle, but the second suffers from the awkwardness of the auxiliary.

Here we have neither the data that justify balance inversion, nor the results that should follow from it. It is due to the false principle of ‘emphasis’ dealt with below in d. and reads as awkwardly as such inversions usually read. The sentence is, no doubt, cumbrous in the uninverted form; but it wants reconstruction, not inversion.

Wrong again, for the same reasons, but not with the same excuse; for the original form is unobjectionable. The emphasis is not on the problem (to find ...), but on the clue to it (much deeper down), which, being emphatic, can maintain its position at the end of the sentence. The compound verb is only a secondary objection: we do not mend matters much by substituting lie for must we go to find.

So is every one is a correct inversion: so is too weak to stand at the end, and at the beginning it is a good enough sign-post to tell us that selfishness is going to be defended. But so is every one selfish is wrong: for if selfish is repeated at all, it is repeated with rhetorical effect, and is strong enough to take care of itself. Our second rule is thus violated; and so is our fourth—the subject does not come at the end.

This time, the compound verb is fatal. ‘So, perhaps, has ...’ would do.

Right or wrong in principle, this does not read comfortably; but that may seem to be due to the cumbrous phrase ‘was there no lack of’, which for practical purposes is a compound verb. That difficulty we can remove without disturbing the accentuation of the sentence: ‘Especially numerous were the speculations as to the character of Miss Hartmann’. This resembles in form our old type ‘Among the guests were ...’, but with the important difference that ‘especially numerous’ is emphatic, and can therefore stand at the end. The inversion is rather explained than justified by the still stronger emphasis on ‘Miss Hartmann’. Sentences in which both subject and predicate are independently emphatic should be avoided, quite apart from the question of inversion: italics are more or less necessary to secure the inferior emphasis, and italics are a confession of weakness.

Not a happy attempt. We notice, for one thing, that the subject does not come at the end; the inversion is not complete. Let us complete it. To do so, we must convey our huge sign-post to the beginning: ‘By this proof ... Thames, was somewhat lightened the provincial panic.’ Worse than ever; is the compound verb to blame? Remove it, and see: ‘In consequence of this proof ... Thames, subsided in some degree the provincial panic’. This is not much better. There is another and a worse flaw: condition number one is not satisfied; we want ‘an emphatic subject that carries in itself the point of the sentence’. Now we must not assume that because ‘provincial’ is italicized, therefore the subject (however emphatic) carries in itself the point of the sentence. What is that point? what imaginary question does the sentence answer? Can it be meant to answer the question ‘What limitations were there upon the comfort derived from the intelligence that the murderer was still in London?’? No; that question could not be asked; we have not yet been told that any comfort at all was derived. The question it answers is ‘What effect did this intelligence produce upon the general panic?’. This question can be asked; for the reader evidently knows that a panic had prevailed, and that the intelligence had come. If, then, we are to use balance inversion, we must so reconstruct the sentence that the words containing the essential answer to this question become the subject; we must change ‘somewhat lightened’ into ‘some alleviation’. ‘From this proof ... Thames, resulted some alleviation of the provincial panic.’ That is the best that inversion will do for us; it is not quite satisfactory, and the reason is that the sentence is made to do too much. When the essential point is subject to an emphatic limitation (an unemphatic one like ‘somewhat’ does not matter), the limitation ought to be conveyed in a separate sentence; otherwise the sentence is overworked, and either shirks its work, with the result of obscurity, or protests by means of italics. We ought therefore to have: ‘From ... resulted some alleviation of the general panic; this, however, was confined to the provinces’. But, except for this incidental fault, the sentence can be mended without inversion: ‘By this proof ... Thames, the provincial panic was somewhat lightened’.

c. Inversion in syntactic clauses.

In clauses introduced by as, than, or a relative (pronoun or adverb), we have only a special case of balance inversion. They differ from the instances considered above in this important respect, that their relation to the preceding words is no longer paratactic, but syntactic, with the result that the sign-post indicating this relation is necessarily placed at the beginning. This will be seen from a comparison of the paratactic and syntactic forms in the following pairs of examples:

Now in each of these sentences there are the same inducements to inversion in the syntactic form as in the paratactic; and added to these is the necessity for placing the sign-post at the beginning. We might expect, therefore, that inversion of syntactic clauses would be particularly common. But (i) We have already seen that inversion does not necessarily follow from the fact that the sign-post is placed at the beginning. And (ii) The verb in as and than clauses will probably, from the nature of the case, be the same as in the preceding clause. If it is in the same mood and tense, it can usually be omitted, unless effective repetition is required, in which case it will go to the end: a change of mood or tense, on the other hand, will often be marked by an auxiliary (itself perhaps compound), which again will usually preclude inversion.

The result is this:

i. Relative clauses, uninfluenced by the position of the sign-post, remain subject to precisely the same conditions as the corresponding paratactic sentences. Thus ‘Among whom were....’ is right, just as ‘Among the guests were....’ was right; ‘Among which would I mention....’ is of course impossible, because the subject does not carry the point; and ‘To which may be attributed....’ is right or wrong, according as the subject is or is not long enough to balance the compound verb.

ii. Inversion of an as or than clause, having become unusual for the reason mentioned above, is almost certain to look either archaic or clumsy; clumsy when the reason for it is apparent, archaic when it is not. The practical rule is this: if you cannot omit the verb, put it at the end; and if you can neither omit it nor put it at the end, reconstruct the sentence.

The verb is preserved to avoid ambiguity. But it should go to the end, especially as it is compound.

Ambiguity cannot fairly be pleaded here; the verb should be omitted.

Either ‘as England did’, or, since the parallel is significant, ‘as England then remained’. Also, ‘as England had’.

The verb must be retained, for the sake of sometimes; but it should go to the end.

‘As few ... have performed them.’

This is like the case considered in b. ‘so is every one selfish’. If improper is repeated with rhetorical effect, there is no need of inversion: if not, it should be left out.

Either ‘would have done’ at the end, or (perhaps better) no verb at all.

Like the indiscriminate use of while, this ungainly as connexion is popular with slovenly writers, and is always aggravated by inversion. ‘All, too, must hope....’

d. Negative inversion, and false ‘emphasis’ inversion.

The connexion here suggested between certain forms of inversion must be taken to represent, not by any means the historical order of development, with which we are not directly concerned, but the order in which a modern writer may be supposed, more or less unconsciously, to adopt them. Starting from an isolated case of necessary inversion, we proceed to extensions of it that seem natural and are sanctioned by modern usage; and from these to other extensions, based probably on a misunderstanding, and producing in modern writers the effect of archaism.

Nor, except when used in conjunction with neither, always stands first; and if the subject appears at all, the sentence is always inverted. This requires no illustration.

On the analogy of nor, many other negative words and phrases are thrown to the beginning of the sentence, and again inversion is the result.

It is at this stage that misconception creeps in. Most of these negative phrases are in themselves emphatic; and from their being placed first (really on the analogy of nor) comes the mistaken idea that they derive emphasis from their position. This paves the way for wholesale inversion: any words, other than the subject, are placed at the beginning; and this not always in order to emphasize the words so placed, but merely to give an impressive effect to the whole. The various steps are marked by the instances that follow. In the first two, inversion may be on the analogy of negatives, or may be designed for emphasis; in the third, emphasis is clearly the motive; and in the rest we have mere impressiveness—not to say mere mannerism.

In view of the rapidly increasing tendency to causeless inversion of all kinds, it is far from certain that this last is intentional satire.

e. Miscellaneous.

(i) In narrated dialogue, the demand for variations of ‘he said’, &c., excuse considerable freedom in the matter of inversion. One or two points, however, may be noticed.

When the subject is a personal pronoun, say is perhaps the only verb with which inversion is advisable. ‘Said I, he, they’, and ‘retorted Jones’: but not ‘enquired I’, ‘rejoined he’, ‘suggested they’.

Compound verbs, as usual, do not lend themselves to inversion:

Inverted said at the beginning is one of the first pitfalls that await the novice who affects sprightliness. It is tolerable, if anywhere, only in light playful verse.

With verbs other than said, this form of inversion is still more decidedly a thing to be left to the poets. ‘Appears Verona’; ‘Rose a nurse of ninety years’; but not

(ii) The inverted conditionals should, had, could, would, were, did, being recommended by brevity and a certain neatness, are all more or less licensed by modern usage. It is worth while, however, to name them in what seems to be their order of merit. Should I, from its frequency, is without taint of archaism; but could and would, and, in a less degree, had, are apt to betray their archaic character by the addition of but (‘would he but consent’); and were and did are felt to be slightly out of date, even without this hint.

(iii) Always, after performing inversion of any kind, the novice should go his rounds, and see that all is shipshape. For want of this precaution, a writer who was no novice, particularly in the matter of inversion, produces such curiosities as these:

We shall venture on removing the comma before ‘though’; but must leave it to connoisseurs in inversion to decide between the rival attractions of ‘disagree with M. Comte though I do’ and ‘disagreeing ... though I am’. ‘Though I do’, in spite of the commas, can scarcely be meant to be parenthetic; that would give (by resolution of the participle) ‘though I disagree with M. Comte, though I do, ....’

a. Occasional.

We have implied in former sections, and shall here take it for granted, that occasional archaism is always a fault, conscious or unconscious. There are, indeed, a few writers—Lamb is one of them—whose uncompromising terms, ‘Love me, love my archaisms’, are generally accepted; but they are taking risks that a novice will do well not to take.

As to unconscious archaism, it might be thought that such a thing could scarcely exist: to employ unconsciously a word that has been familiar, and is so no longer, can happen to few. Yet charitable readers will believe that in the following sentence demiss has slipped unconsciously from a learned pen:

The guilt of such peccadilloes as this may be said to vary inversely as the writer’s erudition; for in this matter the learned may plead ignorance, where the novice knows too well what he is doing. It is conscious archaism that offends, above all the conscious archaisms of the illiterate: the historian’s It should seem, even the essayist’s You shall find, is less odious, though not less deliberate, than the ere, oft, aught, thereanent, I wot, I trow, and similar ornaments, with which amateurs are fond of tricking out their sentences. This is only natural. An educated writer’s choice falls upon archaisms less hackneyed than the amateur’s; he uses them, too, with more discretion, limiting his favourites to a strict allowance, say, of once in three essays. The amateur indulges us with his whole repertoire in a single newspaper letter of twenty or thirty lines, and—what is worse—cannot live up to the splendours of which he is so lavish: charmed with the discovery of some antique order of words, he selects a modern slang phrase to operate upon; he begins a sentence with ofttimes, and ends it with a grammatical blunder; aspires to albeit, and achieves howbeit. Our list begins with the educated specimens, but lower down the reader will find several instances of this fatal incongruity of style; fatal, because the culprit proves himself unworthy of what is worthless. For the vilest of trite archaisms has this latent virtue, that it might be worse; to use it, and by using it to make it worse, is to court derision.

The worst of making a mannerism of this shall is that, after the first two or three times, the reader is certain to see it coming; for its function is nearly always the same—to bring in illustrations of a point already laid down.

Regarded as a counter-irritant to slang, archaism is a failure. Frills is ten times more noticeable for the prim and pompous be.

‘We must not’. Similarly, the modern prose English for if I be, it were, is if I am, it would be.

There, not thither, is the modern form; to it, not thereto; of which, of this, not whereof, hereof; till then, or up to that time, not theretofore. So, in the following examples, except, perhaps, before, though; not save, perchance, ere, albeit.

The use of ere with a gerund is particularly to be avoided.

The writer means albeit; he would have been safer with though.

b. Sustained archaism in narrative and dialogue.

A novelist who places his story in some former age may do so for the sake of a purely superficial variety, without any intention of troubling himself or his readers with temporal colour more than is necessary to avoid glaring absurdities; he is then not concerned with archaism at all. More commonly, however, it is part of his plan to present a living picture of the time of which he writes. When this is the case, he naturally feels bound to shun anachronism not only in externals, but in thought and the expression of thought. Now with regard to the language of his characters, it would be absurd for him to pretend to anything like consistent realism: he probably has no accurate knowledge of the language as his characters would speak it; and if he had this knowledge, and used it, he would be unintelligible to most of his readers, and burdensome to the rest. Accordingly, if he is wise, he will content himself with keeping clear of such modes of expression as are essentially modern and have only modern associations, such as would jar upon the reader’s sense of fitness and destroy the time illusion. He will aim, that is to say, at a certain archaic directness and simplicity; but with the archaic vocabulary, which instead of preserving the illusion only reminds us that there is an illusion to be preserved, he will have little to do. This we may call negative archaism. Esmond is an admirable example of it, and the ‘Dame Gossip’ part of Mr. Meredith’s Amazing Marriage is another. It hardly occurs to us in these books that the language is archaic; it is appropriate, that is all. The same may be said, on the whole, of Treasure Island, and of one or two novels of Besant’s.

Only the novelist who is not wise indulges in positive archaism. He is actuated by the determination to have everything in character at all costs. He does not know very much about old English of any period; very few people do, and those who know most of it would be the last to attempt to write a narrative in it. He gives us, however, all that he knows, without much reference to particular periods; it may not be good ancient English, but, come what may, it shall not be good modern. This, it need scarcely be said, is not fair play: the recreation is all on the writer’s side. Archaism is, no doubt, very seductive to the archaist. Well done (that is, negatively done), it looks easy; and to do it badly is perhaps even easier than it looks. No very considerable stock-in-trade is required; the following will do quite well: Prithee—quotha—perchance—peradventure—i’ faith—sirrah—beshrew me—look ye—sith that—look to it—leave prating—it shall go hard but—I tell you, but—the more part—fair cold water—to me-ward—I am shrewdly afeared—it is like to go stiff with me—y’ are—y’ have—it irks me sorely—benison—staunch—gyves—yarely—this same villain—drink me this—you were better go; to these may be added the indiscriminate use of ‘Nay’ and ‘Now (by the rood, &c.)’; free inversion; and verb terminations in -st and -th. Our list is largely drawn from Stevenson, who, having tried negative archaism with success in Treasure Island, chose to give us a positive specimen in The Black Arrow. How vexatious these reach-me-down archaisms can become, even in the hands of an able writer, will be seen from the following examples of a single trick, all taken from The Black Arrow.

It is like a child with a new toy.

But there is the opposite fault. The judicious archaist, as we have said, will abstain from palpable modernisms, especially from modern slang. The following extracts are taken from an old woman’s reminiscences of days in which a ‘faultless attire’ included ‘half high boots, knee-breeches very tight above the calf (as the fashion was then), a long-tailed cutaway coat, ...’:

Strictly speaking, metaphor occurs as often as we take a word out of its original sphere and apply it to new circumstances. In this sense almost all words can be shown to be metaphorical when they do not bear a physical meaning; for the original meaning of almost all words can be traced back to something physical; in our first sentence above, for instance, there are eight different metaphors. Words had to be found to express mental perceptions, abstract ideas, and complex relations, for which a primitive vocabulary did not provide; and the obvious course was to convey the new idea by means of the nearest physical parallel. The commonest Latin verb for think is a metaphor from vine-pruning; ‘seeing’ of the mind is borrowed from literal sight; ‘pondering’ is metaphorical ‘weighing’. Evidently these metaphors differ in intention and effect from such a phrase as ‘smouldering’ discontent; the former we may call, for want of a better word, ‘natural’ metaphor, as opposed to the latter, which is artificial. The word metaphor as ordinarily used suggests only the artificial kind: but in deciding on the merits or demerits of a metaphorical phrase we are concerned as much with the one class as the other; for in all doubtful cases our first questions will be, what was the writer’s intention in using the metaphor? is it his own, or is it common property? if the latter, did he use it consciously or unconsciously?

This distinction, however, is useful only as leading up to another. We cannot use it directly as a practical test: artificial metaphors, as well as natural ones, often end by becoming a part of ordinary language; when this has happened, there is no telling to which class they belong, and in English the question is complicated by the fact that our metaphorical vocabulary is largely borrowed from Latin in the metaphorical state. Take such a word as explain: its literal meaning is ‘spread out flat’: how are we to say now whether necessity or picturesqueness first prompted its metaphorical use? And the same doubt might arise centuries hence as to the origin of a phrase so obviously artificial to us as ‘glaring inconsistency’.

Our practical distinction will therefore be between conscious or ‘living’ and unconscious or ‘dead’ metaphor, whether natural or artificial in origin: and again, among living metaphors, we shall distinguish between the intentional, which are designed for effect, and the unintentional, which, though still felt to be metaphors, are used merely as a part of the ordinary vocabulary. It may seem at first sight that this classification leaves us where we were: how can we know whether a writer uses a particular metaphor consciously or unconsciously? We cannot know for certain: it is enough if we think that he used it consciously, and know that we should have used it consciously ourselves; experience will tell us how far our perceptions in this respect differ from other people’s. Most readers, we think, will agree in the main with our classification of the following instances; they are taken at random from a couple of pages of the Spectator.

These we should call dead: ‘his views were personal’; ‘carry out his policy’; ‘not acceptable to his colleagues’; ‘the Chancellor proposed’; ‘some grounds for complaint’; ‘refrain from talking about them’; ‘the remission of the Tea-duty’; ‘sound policy’; ‘a speech almost entirely composed of extracts’; ‘reduction of taxation’; ‘discussion’; ‘the low price of Consols’; ‘falls due’; ‘succeeded’; ‘will approach their task’; ‘delivered a speech’; ‘postponing to a future year’. The next are living, but not intentional metaphor; the writer is aware that his phrase is still picturesque in effect, but has not chosen it for that reason: ‘a Protestant atmosphere’; ‘this would leave a margin of £122,000’; ‘the loss of elasticity’ in the Fund; ‘recasting our whole Fiscal system’; ‘to uphold the unity of the Empire’; ‘to strengthen the Exchequer balances’; ‘all dwelt on the grave injury’; ‘his somewhat shattered authority’; ‘the policy of evasion now pursued’; ‘throws new light on the situation’; ‘a gap in our fiscal system’. Intentional metaphors are of course less plentiful: ‘the home-rule motion designed to “draw” Sir Henry’; ‘a dissolving view of General Elections’; ‘this reassuring declaration knocks the bottom out of the plea of urgency’; ‘the scattered remnants of that party might rally after the disastrous defeat’.

One or two general remarks may be made before we proceed to instances. It is scarcely necessary to warn any one against over-indulgence in intentional metaphor; its effects are too apparent. The danger lies rather in the use of live metaphor that is not intentional. The many words and phrases that fall under this class are all convenient; as often as not they are the first that occur, and it is laborious, sometimes impossible, to hit upon an equivalent; the novice will find it worth while, however, to get one whenever he can. We may read a newspaper through without coming upon a single metaphor of this kind that is at all offensive in itself; it is in the aggregate that they offend. ‘Cries aloud for’, ‘drop the curtain on’, ‘goes hand in hand with’, ‘a note of warning’, leaves its impress’, ‘paves the way for’, ‘heralds the advent of’, ‘opens the door to’, are not themselves particularly noisy phrases; but writers who indulge in them generally end by being noisy.

Unintentional metaphor is the source, too, of most actual blunders. Every one is on his guard when his metaphor is intentional; the nonsense that is talked about mixed metaphor, and the celebrity of one or two genuine instances of it that come down to us from the eighteenth century, have had that good effect. There are few obvious faults a novice is more afraid of committing than this of mixed metaphor. His fears are often groundless; many a sentence that might have stood has been altered from a misconception of what mixed metaphor really is. The following points should be observed.

1. If only one of the metaphors is a live one, the confusion is not a confusion for practical purposes.

2. Confusion can only exist between metaphors that are grammatically inseparable; parallel metaphors between which there is no grammatical dependence cannot result in confusion. The novice must beware, however, of being misled either by punctuation or by a parallelism that does not secure grammatical independence. Thus, no amount of punctuation can save the time-honoured example ‘I smell a rat: I see him hovering in the air: ... I will nip him in the bud’. Him is inseparable from the later metaphors, and refers to the rat. But there is no confusion in the following passage; any one of the metaphors can be removed without affecting the grammar:

3. Metaphor within metaphor is dangerous. Here there is a grammatical dependence between the metaphors, and if the combination is unsuitable confusion will result. But combination is one thing, and confusion is another: if the internal metaphor is not inconsistent with the external, there is no confusion, though there may be ugliness. To adapt one of our examples below, ‘The Empire’s butcher (i. e. New Zealand) has not all his eggs in one basket’ is not a confusion, because a metaphorical butcher can have his eggs in one basket as well as any one else. What does lead to confusion is the choice of an internal metaphor applicable not to the words of the external metaphor, but to the literal words for which it is substituted. In the following example, the confusion is doubtless intended.

The swallowing is applicable only to the person metaphorically called a pillar.

4. Confusion of metaphor is sometimes alleged against sentences that contain only one metaphor—a manifest absurdity. These are really cases of a clash between the metaphorical and the non-metaphorical. A striking or original metaphor is apt to appear violent, and a commonplace one impertinent, if not adequately borne out by the rest of the sentence. This we may label ‘unsustained metaphor’. It sometimes produces much the same effect as mixed metaphor; but the remedy for it, as well as the cause, is different. Mixed metaphor is the result of negligence, and can generally be put right by a simple adaptation of the language to whichever metaphor is to be retained. Unsustained metaphor is rather an error of judgement: it is unsustained either because it was difficult to sustain, or because it was not worth sustaining; in either case abandonment is the simplest course.

Here we have four different metaphors; but as they are all dead, there is no real confusion.

Burning and chill are both live metaphors, they are grammatically connected by its, and they are inconsistent; there is therefore confusion.

Extends is usually dead; and if in this case it is living, it is also suitable.

The main or external metaphor is that of an army. Now any metaphor that is applicable to a literal army is also applicable to a metaphorical one: but ‘rally round a nucleus’ is a confusion of metaphor, to whichever it is applied; it requires us to conceive of the army at the same time as animal and vegetable, nucleus being literally the kernel of a nut, and metaphorically a centre about which growth takes place. An army can have a nucleus, but cannot rally round it.

This is quoted in a newspaper as an example of mixed metaphor. It is nothing of the kind: they in New Zealand are detached from the metaphor.

‘A rut’, says the same newspaper, ‘is about the very last thing we should expect to find at sea, despite the fact that it is ploughed’. There is no mention of ruts at sea; the two metaphors are independent. If the speaker had said ‘Moderate churchmen, moving at length from their old moorings, are beginning to lift this question out of the party rut’, we should have had a genuine confusion, the moorings and the rut being then inseparable. Both this sentence and the preceding one, the reader may think, would have been better without the second metaphor; we agree, but it is a question of taste, not of correctness.

This illustrates the danger of deciding hastily on the deadness of a metaphor, however common it may be. Probably any one would have said that the musical idea in incentive had entirely vanished: but the successive attributes keenness and impelling force are too severe a test; the dead metaphor is resuscitated, and a perceptible confusion results.

Another case of resuscitation. Trenchant turns in its grave; and asset, ready to succumb under the violence of athletic reporters, has yet life enough to resent the imputation of a keen edge. As the critic of ‘ruts at sea’ might have observed, the more blunt, the better the assets.

We impose a burden, we apply a stimulus. It looks as if the writer had meant by a short cut to give us both ideas; if so, his guilt is clear; and if we call impose a mere slip in idiom, the confusion is none the less apparent.

These are independent metaphors; and, as thy work is done is applicable to each of them, there is no confusion.

To stave off a brand is not, perhaps, impossible; but we suspect that it would be a waste of energy. The idea of bulk is inseparable from the process of staving off. The metaphor is usually applied to literal abstract nouns, not to metaphorical concretes: ruin and disaster one can suppose to be of a tolerable size; but a metaphorical brand does not present itself to the imagination as any larger than a literal one. We assume that by brand the instrument is meant: the eleventh hour is all too early to set about staving off the mark.

This is a good example of mixed metaphor of the more pronounced type; it differs only in degree from some of those considered above. We suggested that impose a stimulus was perhaps a short cut to the expression of two different metaphors, and the same might be said of staving off the brand. But we shall get a clearer idea of the nature of mixed metaphor if we regard all these as violations of the following simple rule: When a live metaphor (intentional or unintentional) has once been chosen, the words grammatically connected with it must be either (a) recognizable parts of the same metaphorical idea, or one consistent with it, or (b) unmetaphorical, or dead metaphor; literal abstract nouns, for instance, instead of metaphorical concretes. Thus, we shall impose not the stimulus, but either (a) the burden of resistance, or (b) the duty of resistance; and we shall stave off not the ‘brand’ but the ‘ignominy of failure from our military administration’.

But from our remarks in 4 above, it will be clear that (b), though it cannot result in confusion of metaphor, may often leave the metaphor unsustained. Our examples illustrate several common types.

The reader who is uncharitable enough to insist upon the natural history of dilemmas will call this not unsustained metaphor, but a gross confusion; horns cannot be said to screw. We prefer to believe that De Quincey was not thinking of the horns at all; they are a gratuitous metaphorical ornament; dilemma, in English at any rate, is a literal word, and means an argument that presents two undesirable alternatives. The circumstances of a dilemma are, indeed, such as to prompt metaphorical language, but the word itself is incorrigibly literal; we confess as much by clapping horns on its head and making them do the metaphorical work.

This metaphor always requires that the dictator—usually a personified abstract—should be mentioned. ‘Dictated by the importance’.

The opposite fault of over-conscientiousness must also be noticed. Elaborate poetical metaphor has perhaps gone out of fashion; but technical metaphor is apt to be overdone, and something of the same tendency appears in the inexorable working-out of popular catchword metaphors:

We are not photographers enough to hazard a comment on cold print.

‘Rhetorical’ or—to use at once a wider and a more intelligible term—‘significant’ repetition is a valuable element in modern style; used with judgement, it is as truly a good thing as clumsy repetition, the result of negligence, is bad. But there are some writers who, from the fact that all good repetition is intentional, rashly infer that all intentional repetition is good; and others who may be suspected of making repetitions from negligence, and retaining them from a misty idea that to be aware of a thing is to have intended it. Even when the repetition is a part of the writer’s original plan, consideration is necessary before it can be allowed to pass: it is implied in the terms ‘rhetorical’ or significant repetition that the words repeated would ordinarily be either varied or left out; the repetition, that is to say, is more or less abnormal, and whatever is abnormal may be objectionable in a single instance, and is likely to become so if it occurs frequently.

The writers who have most need of repetition, and are most justified in using it, are those whose chief business it is to appeal not to the reader’s emotions, but to his understanding; for, in spite of the term ‘rhetorical’, the object ordinarily is not impressiveness for impressiveness’ sake, but emphasis for the sake of clearness. It may seem, indeed, that a broad distinction ought to be drawn between the rhetorical and the non-rhetorical: they differ in origin and in aim, one being an ancient rhetorical device to secure impressiveness, the other a modern development, called forth by the requirements of popular writers on subjects that demand lucidity; and there is the further difference, that rhetorical repetition often dictates the whole structure of the sentence, whereas the non-rhetorical, in its commonest form, is merely the completion of a sentence that need not have been completed. But in practice the two things become inseparable, and we shall treat them together; only pointing out to the novice that of the two motives, impressiveness and lucidity, the latter is far the more likely to seem justifiable in the reader’s eyes.

We shall illustrate both the good and bad points of repetition almost exclusively from a few pages of Bagehot, one of its most successful exponents, in whom nevertheless it degenerates into mannerism. To a writer who has so much to say that is worth hearing, almost anything can be forgiven that makes for clearness; and in him clearness, vigour, and a certain pleasant rapidity, all result from the free use of repetition. It will be seen that his repetitions are not of the kind properly called rhetorical; it is the spontaneous fullness of a writer who, having a clear point to make, is determined to make it clearly, elegance or no elegance. Yet the growth of mannerism is easily seen in him; the justifiable repetitions are too frequent, and he has some that do not seem justifiable.

Here is a writer who, at any rate, has not the vice of ‘elegant variation’. Most of the possibilities of repetition, for good and for evil, are here represented. As Bagehot himself might have said, ‘we have instances of repetition that are good in themselves; we have instances of repetition that are bad in themselves; and we have instances of repetition that are neither particularly good nor particularly bad in themselves, but that offend simply by recurrence’. The ludicrous appearance presented by our collection as a whole necessarily obscures the merit of individual cases; but if the reader will consider each sentence by itself, he will see that repetition is often a distinct improvement. The point best illustrated here, no doubt, is that it impossible to have too much of a good thing; but it is a good thing for all that. As instances of unjustifiable mannerism, we may select ‘fit to be the favourites ..., and fit to be great orators’; ‘not political communities ..., but political communities ...’; ‘something which he wanted, but something which he did not know that he wanted’; ‘a man like Walpole, or a man like Louis Napoleon’; ‘without reluctance and without cessation’; ‘who did not like ..., who did not like ...’; and ‘without Spanish and without French’. We have mentioned clearness as the ultimate motive for repetition of this kind: in this last sentence, we get not clearness, but obscurity. Any one would suppose that there was some point in the distinction between Spanish and French: there is none; the point is, simply, that languages do not make a statesman. Again, there is sometimes virtue in half-measures: from ‘something which he did not know that he wanted’ remove the first three words, and there remains quite repetition enough. ‘Wild in life and not wild in mind’ is a repetition that is clearly called for; but it is followed by the wholly gratuitous ‘fit ... and fit ...’, and the result is disastrous. Finally, in ‘who did not like ..., who did not like ...’, mannerism gets the upper hand altogether: instead of the appearance of natural vigour that ordinarily characterizes the writer, we have stiff, lumbering artificiality.

Writers like Bagehot do not tend at all to impressive repetition: their motive is always the business-like one of lucidity, though it is sometimes lucidity run mad. Repetition of this kind, not being designed to draw the reader’s attention to itself, wears much better in practice than the more pronounced types of rhetorical repetition. The latter should be used very sparingly. As the spontaneous expression of strong feeling in the writer, it is sometimes justified by circumstances: employed as a deliberate artifice to impress the reader, it is likely to be frigid, and to fail in its object; and the term ‘rhetorical’ should remind us in either case that what may be spoken effectively will not always bear the test of writing.

Rhetorical repetition, when it is clearly distinguishable from the non-rhetorical, is too obvious to require much illustration. Of the three instances given, the last is an excellent test case for the principle that ‘whatever is intentional is good’.

a. Some more trite phrases.

The worn-out phrases considered in a former section were of a humorous tendency: we may add here some expressions of another kind, all of them calculated in one way or another to save the writer trouble; the trouble of description, or of producing statistics, or of thinking what he means. Such phrases naturally die hard; even ‘more easily imagined than described’ still survives the rough handling it has met with, and flourishes in writers of a certain class. ‘Depend upon it’, ‘you may take my word for it’, ‘in a vast majority of cases’, ‘no thinking man will believe’, ‘all candid judges must surely agree’, ‘it would be a slaying of the slain’, ‘I am old-fashioned enough to think’, are all apt to damage the cause they advocate.

The shrill formula ‘It stands to reason’ is one of the worst offenders. Originally harmless, and still no doubt often used in quite rational contexts, the phrase has somehow got a bad name for prefacing fallacies and for begging questions; it lacks the delicious candour of its feminine equivalent—‘Kindly allow me to know best’—, but appeals perhaps not less irresistibly to the generosity of an opponent. Apart from this, there is a correct and an incorrect use of the words. It is of course the conclusion drawn from certain premisses that stands to reason; the premisses do not stand to reason; they are assumed to be a matter of common knowledge, and ought to be distinguished from the conclusion by if or a causal participle, not co-ordinated with it by and.

Just as ‘stands to reason’ is not an argument, but an invitation to believe, ‘the worthy Major’ not amusing, but an invitation to smile, so the sentimental or sensational novelist has his special vocabulary of the impressive, the tender, the tragic, and the horrible. One or two of the more obvious catch-phrases may be quoted. In the ‘strong man’ of fiction the reader may have observed a growing tendency to ‘sob like a child’; the right-minded hero to whom temptation comes decides, with archaic rectitude, that he ‘will not do this thing’; the villain, taught by incessant ridicule to abstain from ‘muffled curses’, finds a vent in ‘discordant laughs, that somehow jarred unpleasantly upon my nerves’; this laugh, mutatis mutandis (‘cruel little laugh, that somehow ...’), he shares with the heroine, who for her exclusive perquisite has ‘this man who had somehow come into her life’. Somehow and half-dazed are invaluable for throwing a mysterious glamour over situations and characters that shun the broad daylight of common sense.

b. Elementary irony.

A well-known novelist speaks of the resentment that children feel against those elders who insist upon addressing them in a jocular tone, as if serious conversation between the two were out of the question. Irony is largely open to the same objection: the writer who uses it is taking our intellectual measure; he forgets our ex officio perfection in wisdom. Theoretically, indeed, the reader is admitted to the author’s confidence; he is not the corpus vile on which experiment is made: that, however, is scarcely more convincing than the two-edged formula ‘present company excepted’. For minute, detailed illustration of truths that have had the misfortune to become commonplaces without making their due impression, sustained irony has its legitimate use: tired of being told, and shown by direct methods, that only the virtuous man is admirable, we are glad enough to go off with Fielding on a brisk reductio ad absurdum: ‘for if not, let some other kind of man be admirable; as Jonathan Wild’. But the reductio process should be kept for emergencies, as Euclid kept it, with whom it is a confession that direct methods are not available. The isolated snatches of irony quoted below have no such justification: they are for ornament, not for utility; and it is a kind of ornament that is peculiarly un-English—a way of shrugging one’s shoulders in print.

Pleasant is put for unpleasant because the latter seemed dull and unnecessary; the writer should have taken the hint, and put nothing at all.

The climax is reached by those pessimists who, regarding the reader’s case as desperate, assist him with punctuation, italics, and the like:

c. Superlatives without the.

The omission of the with superlatives is limited by ordinary prose usage to (1) Superlatives after a possessive: ‘Your best plan’. (2) Superlatives with most: ‘in most distressing circumstances’, but not ‘in saddest circumstances’. (3) Superlatives in apposition, followed by of: ‘I took refuge with X., kindliest of hosts’; ‘We are now at Weymouth, dingiest of decayed watering-places’. Many writers of the present day affect the omission of the in all cases where the superlative only means very. No harm will be done if they eventually have their way: in the meantime, the omission of the with inflected superlatives has the appearance of gross mannerism.

d. Cheap originality.

Just as ‘elegant variation’ is generally a worse fault than monotony, so the avoidance of trite phrases is sometimes worse than triteness itself. Children have been known to satisfy an early thirst for notoriety by merely turning their coats inside out; and ‘distinction’ of style has been secured by some writers on the still easier terms of writing a common expression backwards. By this simplest of all possible expedients, ‘wear and tear’ ceases to be English, and becomes Carlylese, and Emerson acquires an exclusive property (so at least one hopes) in ‘nothing or little’. The novice need scarcely be warned against infringing these writers’ patents; it would be as unpardonable as stealing the idea of a machine for converting clean knives into dirty ones. Hackneyed phrases become hackneyed because they are useful, in the first instance; but they derive a new efficiency from the very fact that they are hackneyed. Their precise form grows to be an essential part of the idea they convey, and all that a writer effects by turning such a phrase backwards, or otherwise tampering with it, is to give us our triteness at secondhand; we are put to the trouble of translating ‘tear and wear’, only to arrive at our old friend ‘wear and tear’, hackneyed as ever.

CHAPTER IV PUNCTUATION — The Colon — Dashes — Hyphens — Quotation Marks — FOOTNOTES:

In this chapter we shall adhere generally to our plan of not giving systematic positive directions, or attempting to cover all ground familiar and unfamiliar, important or not, but drawing attention only to the most prevalent mistakes. On so technical a subject, however, a few preliminary remarks may be made; and to those readers who would prefer a systematic treatise Beadnell’s Spelling and Punctuation (Wyman’s Technical Series, Menken, 2/6) may be recommended. We shall refer to it occasionally in what follows; and the examples to which —B. is attached instead of an author’s name are taken from it; these are all given in Beadnell (unless the contrary is stated) as examples of correct punctuation. It should be added that the book is written rather from the compositor’s than from the author’s point of view, and illustrates the compositor’s natural weaknesses; it is more important to him, for instance, that a page should not be unsightly (the unsightliness being quite imaginary, and the result of professional conservatism) than that quotation marks and stops, or dashes and stops, should be arranged in their true significant order; but, as the right and unsightly is candidly given as well as the wrong and beautiful, this does not matter; the student can take his choice.

We shall begin by explaining how it is that punctuation is a difficult matter, and worth a writer’s serious attention. There are only six stops, comma, semicolon, colon, full stop, question mark, exclamation mark; or, with the dash, seven. The work of three of them, full stop, question, exclamation, is so clear that mistakes about their use can hardly occur without gross carelessness; and it might be thought that with the four thus left it ought to be a very simple matter to exhaust all possibilities in a brief code of rules. It is not so, however. Apart from temporary disturbing causes—of which two now operative are (1) the gradual disappearance of the colon in its old use with the decay of formal periodic arrangement, and (2) the encroachments of the dash as a saver of trouble and an exponent of emotion—there are also permanent difficulties.

Before mentioning these we observe that the four stops in the strictest acceptation of the word (,) (;) (:) (.)—for (!) and (?) are tones rather than stops—form a series (it might be expressed also by 1, 2, 3, 4), each member of which directs us to pause for so many units of time before proceeding. There is essentially nothing but a quantitative time relation between them.

The first difficulty is that this single distinction has to convey to the reader differences of more than one kind, and not commensurable; it has to do both logical and rhetorical work. Its logical work is helping to make clear the grammatical relations between parts of a sentence or paragraph and the whole or other parts: its rhetorical work is contributing to emphasis, heightening effect, and regulating pace. It is in vain that Beadnell lays it down: ‘The variation of pause between the words of the same thought is a matter of rhetoric and feeling, but punctuation depends entirely upon the variation of relations—upon logical and grammatical principles’. The difference between these two:

is in logic nothing; but in rhetoric it is the difference between matter-of-fact statement and indignant statement: a strap, we are to understand from the comma, is a barbarous instrument.

Again, in the two following examples, so far as logic goes, commas would be used in both, or semicolons in both. But the writer of the second desires to be slow, staccato, and impressive: the writer of the first desires to be rapid and flowing, or rather, perhaps, does not desire to be anything other than natural.

The difference is rhetorical, not logical. It is true, however, that modern printers make an effort to be guided by logic or grammar alone; it is impossible for them to succeed entirely; but any one who will look at an Elizabethan book with the original stopping will see how far they have moved: the old stopping was frankly to guide the voice in reading aloud, while the modern is mainly to guide the mind in seeing through the grammatical construction.

A perfect system of punctuation, then, that should be exact and uniform, would require separate rhetorical and logical notations in the first place. Such a system is not to be desired; the point is only that, without it, usage must fluctuate according as one element is allowed to interfere with the other. But a second difficulty remains, even if we assume that rhetoric could be eliminated altogether. Our stop series, as explained above, provides us with four degrees; but the degrees of closeness and remoteness between the members of sentence or paragraph are at the least ten times as many. It is easy to show that the comma, even in its purely logical function, has not one, but many tasks to do, which differ greatly in importance. Take the three examples:

The removal of the comma after learned makes so little difference that it is an open question among compositors whether it should be used or not.

With the commas, the criminal is necessarily a certain person already known to us: without them, we can only suppose a past state of society to be described, in which all traitors were ashamed of themselves—a difference of some importance.

Omission of the comma after appointed gives us two persons instead of one, and entirely changes the meaning, making the central words into, what they could not possibly be with the comma, an absolute construction.

These commas, that is, have very different values; many intermediate degrees might be added. Similarly the semicolon often separates grammatically complete sentences, but often also the mere items of a list, and between these extremes it marks other degrees of separation. A perfect system for the merely logical part of punctuation, then, would require some scores of stops instead of four. This again is not a thing to be desired; how little, is clear from the fact that one of our scanty supply, the colon, is now practically disused as a member of the series, and turned on to useful work at certain odd jobs that will be mentioned later. A series of stops that should really represent all gradations might perhaps be worked by here and there a writer consistently with himself; but to persuade all writers to observe the same distinctions would be hopeless.

A third difficulty is this: not only must many tasks be performed by one stop; the same task is necessarily performed by different stops according to circumstances; as if polygamy were not bad enough, it is complicated by an admixture of polyandry. We have already given two sentences of nearly similar pattern, one of which had its parts separated by commas, the other by semicolons, and we remarked that the difference was there accounted for by the intrusion of rhetoric. But the same thing occurs even when logic or grammar (it should be explained that grammar is sometimes defined as logic applied to speech, so that for our purposes the two are synonymous) is free from the disturbing influence; or when that influence acts directly, not on the stop itself that is in question, but only on one of its neighbours. To illustrate the first case, when the stops are not affected by rhetoric, but depend on grammar alone, we may take a short sentence as a nucleus, elaborate it by successive additions, and observe how a particular stop has to go on increasing its power, though it continues to serve only the same purpose, because it must keep its predominance.

The function of the comma is to mark the division between the subordinate and the main clauses.

The semicolon is doing now exactly what the comma did before; but, as commas have intruded into the clause to do the humble yet necessary work of marking two appositions, the original comma has to dignify its relatively more important office by converting itself into a semicolon.

The new insertion is also an apposition, like the former ones; but, as it contains commas within itself, it must be raised above their level by being allowed a semicolon to part it from them. The previous semicolon, still having the same supreme task to do, and challenged by an upstart rival, has nothing for it but to change the regal for the imperial crown, and become a colon. A careful observer will now object that, on these principles, our new insertion ought to have had an internal semicolon, to differentiate the subordinate clause, as if, &c., from the mere enumeration commas that precede: in which case the semi-colon after caprice should be raised to a colon; and then what is the newly created emperor to do? There is no papal tiara for him to assume, the full stop being confined to the independent sentence. The objection is quite just, and shows how soon the powers of the four stops are exhausted if relentlessly worked. But we are concerned only to notice that the effect of stops, even logically considered, is relative, not absolute. It is also true that many modern writers, if they put down a sentence like this, would be satisfied with using commas throughout; the old-fashioned air of the colon will hardly escape notice. But the whole arrangement is according to the compositor’s art in its severer form.

A specimen of the merely indirect action of rhetoric may be more shortly disposed of. In a sentence already quoted—

suppose the writer to have preferred for impressive effect, as we said he might have, to use semicolons instead of commas. The immediate result of that would be that what before could be left to the reader to do for himself (i. e., the supplying of the words have sought knowledge in each member) will in presence of the semicolon require to be done to the eye by commas, and the sentence will run:

But, lest we should be thought too faithful followers of the logicians, we will now assume that our point has been sufficiently proved: the difficulties of punctuation, owing to the interaction of different purposes, and the inadequacy of the instruments, are formidable enough to be worth grappling with.

We shall now only make three general remarks before proceeding to details. The first is implied in what has been already said: the work of punctuation is mainly to show, or hint at, the grammatical relation between words, phrases, clauses, and sentences; but it must not be forgotten that stops also serve to regulate pace, to throw emphasis on particular words and give them significance, and to indicate tone. These effects are subordinate, and must not be allowed to conflict with the main object; but as the grammatical relation may often be shown in more than one way, that way can be chosen which serves another purpose best.

Secondly, it is a sound principle that as few stops should be used as will do the work. There is a theory that scientific or philosophic matter should be punctuated very fully and exactly, whereas mere literary work can do with a much looser system. This is a mistake, except so far as scientific and philosophic writers may desire to give an impressive effect by retarding the pace; that is legitimate; but otherwise, all that is printed should have as many stops as help the reader, and not more. A resolution to put in all the stops that can be correctly used is very apt to result in the appearance of some that can only be used incorrectly; some of our quotations from Huxley and Mr. Balfour may be thought to illustrate this. And whereas slight stopping may venture on small irregularities, full stopping that is incorrect is also unpardonable. The objection to full stopping that is correct is the discomfort inflicted upon readers, who are perpetually being checked like a horse with a fidgety driver.

Thirdly, every one should make up his mind not to depend on his stops. They are to be regarded as devices, not for saving him the trouble of putting his words into the order that naturally gives the required meaning, but for saving his reader the moment or two that would sometimes, without them, be necessarily spent on reading the sentence twice over, once to catch the general arrangement, and again for the details. It may almost be said that what reads wrongly if the stops are removed is radically bad; stops are not to alter meaning, but merely to show it up. Those who are learning to write should make a practice of putting down all they want to say without stops first. What then, on reading over, naturally arranges itself contrary to the intention should be not punctuated, but altered; and the stops should be as few as possible, consistently with the recognized rules. At this point those rules should follow; but adequately explained and illustrated, they would require a volume; and we can only speak of common abuses and transgressions of them.

First comes what may be called for short the spot-plague—the tendency to make full-stops do all the work. The comma, most important, if slightest, of all stops, cannot indeed be got rid of, though even for that the full-stop is substituted when possible; but the semicolon is now as much avoided by many writers as the colon (in its old use) by most. With the semicolon go most of the conjunctions. Now there is something to be said for the change, or the two changes: the old-fashioned period, or long complex sentence, carefully worked out with a view to symmetry, balance, and degrees of subordination, though it has a dignity of its own, is formal, stiff, and sometimes frigid; the modern newspaper vice of long sentences either rambling or involved (far commoner in newspapers than the spot-plague) is inexpressibly wearisome and exasperating. Simplification is therefore desirable. But journalists now and then, and writers with more literary ambition than ability generally, overdo the thing till it becomes an affectation; it is then little different from Victor Hugo’s device of making every sentence a paragraph, and our last state is worse than our first. Patronizing archness, sham ingenuousness, spasmodic interruption, scrappy argument, dry monotony, are some of the resulting impressions. We shall have to trouble the reader with at least one rather long specimen; the spot-plague in its less virulent form, that is, when it is caused not by pretentiousness or bad taste, but merely by desire to escape from the period, does not declare itself very rapidly. What follows is a third or so of a literary review, of which the whole is in exactly the same style, and which might have been quoted entire for the same purpose. It will be seen that it shows twenty full-stops to one semicolon and no colons. Further, between no two of the twenty sentences is there a conjunction.

The reader’s experience has probably been that the constant fresh starts are at first inspiriting, that about half-way he has had quite enough of the novelty, and that he is intensely grateful, when the solitary semicolon comes into sight, for a momentary lapse into ordinary gentle progress. Writers like this may almost be suspected of taking literally a summary piece of advice that we have lately seen in a book on English composition: Never use a semicolon when you can employ a full-stop. Beadnell lays down a law that at first sight seems to amount to the same thing: The notion of parting short independent sentences otherwise than by a full-stop, rests upon no rational foundation, and leads to endless perplexities. But his practice clears him of the imputation: he is saved by the ambiguity of the word independent. There are grammatical dependence, and dependence of thought. Of all those ‘little hard round unconnected things’, in the Times review, that ‘seem to come upon one as shot would descend from a shot-making tower’ (Sir Arthur Helps), hardly one is not dependent on its neighbours in the more liberal sense, though each is a complete sentence and independent in grammar. Now one important use of stops is to express the degrees of thought dependence. A style that groups several complete sentences together, by the use of semicolons, because they are more closely connected in thought, is far more restful and easy—for the reader, that is—than the style that leaves him to do the grouping for himself; and yet it is free from the formality of the period, which consists, not of grammatically independent sentences, but of a main sentence with many subordinate clauses. We have not space for a long example of the group system rightly applied; most good modern writers free from the craving to be up to date will supply them on every page; but a very short quotation may serve to emphasize the difference between group and spot-plague principles. The essence of the latter is that almost the only stops used are full-stops and commas, that conjunctions are rare, and that when a conjunction does occur the comma is generally used, not the full-stop. What naturally follows is an arrangement of this kind:

(To save trouble, let it be stated that the sheil is a dependency of the Home Farm, and not contrasted with or opposed to it.) Here there are three grammatically independent sentences, between the two latter of which the conjunction but is inserted. It follows from spot-plague principles that there will be a full-stop at the end of the first, and a comma at the end of the second. With the group system it is not so simple a matter; before we can place the stops, we have to inquire how the three sentences are connected in thought. It then appears that the friendliness of the grieve is mentioned to account for the sheil’s being at disposal; that is, there is a close connexion, though no conjunction, between the first and the second sentences. Further, the birds in the bush of the third sentence are contrasted, not with the second sentence’s friendliness, but with the first sentence’s bird in the hand (which, however, is accounted for by the second sentence’s friendliness). To group rightly, then, we must take care, quite reversing the author’s punctuation, that the first and second are separated by a stop of less power than that which separates the third from them. Comma, semicolon, would do it, if the former were sufficient between two grammatically independent sentences not joined by a conjunction; it obviously is not sufficient here (though in some such pairs it might be); so, instead of comma, semicolon, we must use semicolon, full-stop; and the sentence will run, with its true meaning much more clearly given:

The group system gives more trouble to the writer or compositor, and less to the reader; the compositor cannot be expected to like it, if the burden falls on him; inferior writers cannot be expected to choose it either, perhaps; but the good writers who do choose it no doubt find that after a short time the work comes to do itself by instinct.

We need now only add two or three short specimens, worse, though from their shortness less remarkable, than the Times extract. They are not specially selected as bad; but it may be hoped that by their juxtaposition they may have some deterrent effect.

The last sentence here is a mere comment on what is itself only an appendage, the clause introduced by especially; it has therefore no right to the dignity of a separate sentence. But it can hardly be mended without some alteration of words as well as stops; for instance, put a semicolon after suitors, write moreover for especially as, and put only a comma after skinflint; the right proportion would then be secured.

The spot-plague, as we have shown, sometimes results in illogicality; it need not do so, however; when it does, the fault lies with the person who, accepting its principles, does not arrange his sentences to suit them. It is a new-fashioned and, in our opinion, unpleasant system, but quite compatible with correctness.

Over-stopping, to which we now proceed, is on the contrary old-fashioned; but it is equally compatible with correctness. Though old-fashioned, it still lingers obstinately enough to make some slight protest desirable; the superstition that every possible stop should be inserted in scientific and other such writing misleads compositors, and their example affects literary authors who have not much ear. Any one who finds himself putting down several commas close to one another should reflect that he is making himself disagreeable, and question his conscience, as severely as we ought to do about disagreeable conduct in real life, whether it is necessary. He will find that the parenthetic or emphatic effect given to an adverbial phrase by putting a comma at each end of it is often of no value whatever to his meaning; in other words, that he can make himself agreeable by merely putting off a certain pompous solemnity; erasing a pair of commas may make the difference in writing that is made in conversation by a change of tone from the didactic to the courteous. Sometimes the abundance of commas is not so easily reduced; a change in the order of words, the omission of a needless adverb or conjunction, even the recasting of a sentence, may be necessary. But it is a safe statement that a gathering of commas (except on certain lawful occasions, as in a list) is a suspicious circumstance. The sentence should at least be read aloud, and if it halts or jolts some change or other should be made.

None of the last three commas is wanted; those round at last are very unpleasant, and they at least should be omitted.

Perhaps can do very well without commas.

Indeed can do without commas, if it cannot itself be done without.

If just is omitted, there need be no commas round occasionally. There may be a value in just; but hardly enough to compensate for the cruel jerking at the bit to which the poor reader is subjected by a remorseless driver.

The comma after thus is nothing if not pompous. And another can be got rid of by putting it may have been before judged by modern lights.

Omit namely and its commas.

A good example of the warning value of commas. None of these can be dispensed with, since there are no less than three parenthetic qualifications to the sentence. But the crowd of commas ought to have told the writer how bad his sentence was; it is like an obstacle race. It should begin, It is true that ..., which disposes of one obstacle. As I have said can be given a separate sentence afterwards—So much has been said before.

Putting commas round apparently amounts to the insertion of a further clause, such as, Though you would not think they could be such fools. But what the precise contents of the further clause may be is problematic. At any rate, a writer should not invite us to read between the lines unless he is sure of two things: what he wants to be read there; and that we are likely to be willing and able readers of it. The same is true of many words that are half adverbs and half conjunctions, like therefore. We have the right to comma them off if we like; but, unless it is done with a definite purpose, it produces perplexity as well as heaviness. In the first of the next two examples, there is no need whatever for the commas. In the second, the motive is clear: having the choice between commas and no commas, the reporter uses them because he so secures a pause after he, and gives the word that emphasis which in the speech as delivered doubtless made the I that it represents equivalent to I for my part.

After for and and beginning a sentence commas are often used that are hardly even correct. It may be suspected that writers allow themselves to be deceived by the false analogy of sentences in which the and or for is immediately followed by a subordinate clause or phrase that has a right to its two commas. When there is no such interruption, the only possible plea for the comma is that it is not logical but rhetorical, and conveys some archness or other special significance such as is hardly to be found in our two examples:

We shall class here also, assuming for the present that the rhetorical plea may be allowed even when there is no logical justification for a stop, two sentences in which the copula is, standing between subject and complement, has commas on each side of it. Impressiveness is what is aimed at; it seems to us a tawdry device for giving one’s sentence an ex cathedra air:

Many other kinds of over-stopping might be illustrated; but we have intentionally confined ourselves here to specimens in which grammatical considerations do not arise, and the sentence is equally correct whether the stops are inserted or not. Sentences in which over-stopping outrages grammar more or less decidedly will be incidentally treated later on. Meanwhile we make the general remark that ungrammatical insertion of stops is a high crime and misdemeanour, whereas ungrammatical omission of them is often venial, and in some cases even desirable. Nevertheless the over-stopping that offends against nothing but taste has its counterpart in under-stopping of the same sort. And it must be added that nothing so easily exposes a writer to the suspicion of being uneducated as omission of commas against nearly universal custom. In the examples that follow, every one will see at the first glance where commas are wanting. When it is remembered that, as we have implied, an author has the right to select the degree of intensity, or scale, of his punctuation, it can hardly be said that grammar actually demands any stops in these sentences taken by themselves. Yet the effect, unless we choose to assume misprints, as we naturally do in isolated cases, is horrible.

Although we are, when we turn from taste to grammar, on slightly firmer ground, it will be seen that there are many debatable questions; and we shall have to use some technical terms. As usual, only those points will be attended to which our observation has shown to be important.

1. The substantival clause.

Subordinate clauses are sentences containing a subject and predicate, but serving the purpose in the main sentence (to which they are sometimes joined by a subordinating conjunction or relative pronoun, but sometimes without any separate and visible link) of single words, namely, of noun, adjective, or adverb; they are called respectively substantival, adjectival, or adverbial clauses. Examples:

Substantival. He asked what I should do. (my plan, noun)

Adjectival. The man who acts honestly is respected. (honest, adjective)

Adverbial. I shall see you when the sun next rises. (to-morrow, adverb)

Now there is no rule that subordinate clauses must be separated from the main sentence by a stop; that depends on whether they are essential parts of the proposition (when stops are generally wrong), or more or less separable accidents (when commas are more or less required). But what we wish to draw attention to is a distinction in this respect, very generally disregarded, between the substantival clause and the two other kinds. When the others are omitted, though the desired meaning may be spoilt, the grammar generally remains uninjured; a complete, though not perhaps valuable sentence is left. The man is respected, I shall see you, are as much sentences alone as they were with the adjectival and adverbial clauses. With substantival clauses this is seldom true; they are usually the subjects, objects, or complements, of the verbs, that is, are grammatically essential. He asked is meaningless by itself. (Even if the point is that he asked and did not answer, things, or something, has to be supplied in thought.) Now it is a principle, not without exceptions, but generally sound, that the subject, object, or complement, is not to be separated from its verb even by a comma (though two commas belonging to an inserted parenthetic clause or phrase or word may intervene). It follows that there is no logical or grammatical justification, though there may be a rhetorical one, for the comma so frequently placed before the that of an indirect statement. Our own opinion (which is, however, contrary to the practice of most compositors) is that this should always be omitted except when the writer has a very distinct reason for producing rhetorical impressiveness by an unusual pause. Some very ugly overstopping would thus be avoided.

The comma before whether in the next is actually misleading; we are tempted to take as adverbial what is really a substantival clause, object to the verbal noun indifference:

Grammar, however, would afford some justification for distinguishing between the substantival clause as subject, object, or complement, and the substantival clause in apposition with one of these. Though there should decidedly be no comma in He said that ..., it is strictly defensible in It is said, that.... The that-clause in the latter is explanatory of, and in apposition with, it; and the ordinary sign of apposition is a comma. Similarly, My opinion is that: It is my opinion, that. But as there seems to be no value whatever in the distinction, our advice is to do without the comma in all ordinary cases of either kind. A useful and reasonable exception is made in some manuals; for instance, in Bigelow’s Manual of Punctuation we read: ‘Clauses like “It is said”, introducing several propositions or quotations, each preceded by the word that, should have a comma before the first that. But if a single proposition or quotation only is given, no comma is necessary. Example:

Anything that shows the reader what he is to expect, and so saves him the trouble of coming back to revise his first impressions, is desirable if there is no strong reason against it.

A more important distinction is this: He said, &c., may have for its object, and It is said, &c., for its (virtual) subject, either the actual words said, or a slight rearrangement of them (not necessarily to the eye, but at least to the mind), which makes them more clearly part of the grammatical construction, and turns them into true subordinate clauses. Thus He told her, You are in danger may be kept, but is usually altered to He told her that she was in danger, or to He told her she was in danger. In the first, You are in danger is not properly a subordinate clause, but a sentence, which may be said to be in apposition with these words understood. In the second and third alike, the altered words are a subordinate substantival clause, the object to told. It follows that when the actual words are given as such (this is sometimes only to be known by the tone: compare I tell you, I will come, and I tell you I will come), a comma should be inserted; whereas, when they are meant as mere reported or indirect speech, it should be omitted. Actual words given as such should also be begun with a capital letter; and if they consist of a compound sentence, or of several sentences, a comma will not suffice for their introduction; a colon, a colon and dash, or a full stop, with quotation marks always in the last case, and usually in the others, will be necessary; but these are distinctions that need not be considered here in detail.

Further, it must be remembered that substantival clauses include indirect questions as well as indirect statements, and that the same rules will apply to them. The two following examples are very badly stopped:

Accommodation of the stops to the words would give:

And accommodation of the words to the stops would give:

The two correct alternatives here are similarly:

As the sentences stood originally, we get in the Carlyle a most theatrical, and in the Times a most slovenly effect.

2. The verb and its subject, object, or complement.

Our argument against the common practice of placing a comma before substantival that-clauses and others like them was, in brief: This sort of that-clause is simply equivalent to a noun; that noun is, with few exceptions, the subject, object, or complement, to a verb; and between things so closely and essentially connected as the verb and any of these no stop should intervene (unless for very strong and special rhetorical reasons). This last principle, that the verb and its essential belongings must not be parted, was merely assumed. We think it will be granted by any one who reads the next two examples. It is felt at once that a writer who will break the principle with so little excuse as here will shrink from nothing.

Very little better than these, but each with some perceptible motive, are the next six:

In each of these the comma acts as an accent upon him, and is purely rhetorical and illogical.

Here the comma guards us from taking you are together. We have already said that this device is illegitimate. Such sentences should be recast; for instance, Women like you are seldom, &c.

Loveliness, maidens, view, the strict subjects, have adjectival phrases attached after them. The temptation to insert the comma is comprehensible, but slight, and should have been resisted.

In the three that come next, the considerable length of the subject, it must be admitted, makes a comma comforting; it gives us a sort of assurance that we have kept our hold on the sentence. It is illogical, however, and, owing to the importance of not dividing subject from verb, unpleasantly illogical. In each case the comfort would be equally effective if it were legitimized by the insertion of a comma before as well as after the clause or phrase at the end of which the present comma stands. The extra commas would be after earth, victims, Schleiden.

It may be said that it is ‘fudging’ to find an excuse, as we have proposed to do, for a stop that we mean really to do something different from its ostensible work. But the answer is that with few tools and many tasks to do much fudging is in fact necessary.

A special form of this, in protest against which we shall give five examples, each from a different well-known author, is when the subject includes and ends with a defining relative clause, after which an illogical comma is placed. As the relative clause is of the defining kind (a phrase that has been explained[12]), it is practically impossible to fudge in these sentences by putting a comma before the relative pronoun. Even in the first sentence the length of the relative clause is no sufficient excuse; and in all the others we should abolish the comma without hesitation.

In the two sentences that now follow from Mr. Morley, the offending comma of the first parts centre, which is what grammarians call the oblique complement, from its verb made; the offending comma of the second parts the direct object groups from its verb drew. Every one will allow that the sentences are clumsy; most people will allow that the commas are illogical. As for us, we do not say that, if the words are to be kept as they are, the commas should be omitted; but we do say that a good writer, when he found himself reduced to illogical commas, should have taken the trouble to rearrange his words.

It may be added, by way of concluding this section, that the insertion of a comma in the middle of an absolute construction, which is capable, as was shown in the sentence about Colonel Hutchinson and the governor, of having very bad results indeed, is only a particular instance and reductio ad absurdum of inserting a comma between subject and verb. The comma in the absolute construction is so recognized a trap that it might have been thought needless to mention it; the following instances, however, will show that a warning is even now necessary.

3. The adjectival clause.

This, strictly speaking, does the work of an adjective in the sentence. It usually begins with a relative pronoun, but sometimes with a relative adverb. The man who does not breathe dies, is equivalent to The unbreathing man dies. The place where we stand is holy ground, is equivalent to This place is holy ground. But we shall include under the phrase all clauses that begin with a relative, though some relative clauses are not adjectival, because a division of all into defining clauses on the one hand, and non-defining or commenting on the other, is more easily intelligible than the division into adjectival and non-adjectival. This distinction is more fully gone into in the chapter on Syntax, where it is suggested that that, when possible, is the appropriate relative for defining, and which for non-defining clauses. That, however, is a debatable point, and quite apart from the question of stopping that arises here. Examples of the two types are:

(Defining) The river that (which) runs through London is turbid.

(Commenting) The Thames, which runs through London, is turbid.

It will be seen that in the first the relative clause is an answer to the imaginary question, ‘Which river?’; that is, it defines the noun to which it belongs. In the second, such a question as ‘Which Thames?’ is hardly conceivable; the relative clause gives us a piece of extra and non-essential information, an independent comment. The two types are not always so easily distinguished as in these examples constructed for the purpose. What we wish here to say is that it would contribute much to clearness of style if writers would always make up their minds whether they intend a definition or a comment, and would invariably use no commas with a defining clause, and two commas with a non-defining. All the examples that follow are in our opinion wrong. The first three are of defining relative clauses wrongly preceded by commas; the second three of commenting relative clauses wrongly not preceded by commas. The last of all there may be a doubt about. If the long clause beginning with which is intended merely to show how great the weariness is, and which is practically equivalent to so great that, it may be called a defining clause, and the omission of the comma is right. But if the which really acts as a mere connexion to introduce a new fact that the correspondent wishes to record, the clause is non-defining, and the comma ought according to our rule to be inserted before it.

4. The adverb, adverbial phrase, and adverbial clause.

In writing of substantival and adjectival clauses, our appeal was for more logical precision than is usual. We said that the comma habitual before substantival clauses was in most cases unjustifiable, and should be omitted even at the cost of occasional slight discomfort. We said that with one division of adjectival, or rather relative clauses, commas should always be used, and with another they should always be omitted. With the adverbial clauses, phrases, and words, on the other hand, our appeal is on the whole for less precision; we recommend that less precision should be aimed at, at least, though more attained, than at present. Certain kinds of laxity here are not merely venial, but laudable: certain other kinds are damning evidence of carelessness or bad taste or bad education. It is not here a mere matter of choosing between one right and one wrong way; there are many degrees.

Now is an adverb; in the house is usually an adverbial phrase; if I know it is an adverbial clause. Logic and grammar never prohibit the separating of any such expressions from the rest of their sentence—by two commas if they stand in the middle of it, by one if they begin or end it. But use of the commas tends, especially with a single word, but also with a phrase or clause, though in inverse proportion to its length, to modify the meaning. I cannot do it now means no more than it says: I cannot do it, now conveys a further assurance that the speaker would have been delighted to do it yesterday or will be quite willing tomorrow. This distinction, generally recognized with the single word, applies also to clauses; and writers of judgement should take the fullest freedom in such matters, allowing no superstition about ‘subordinate clauses’ to force upon them commas that they feel to be needless, but inclining always when in doubt to spare readers the jerkiness of overstopping. It is a question for rhetoric alone, not for logic, so long as the proper allowance of commas, if any, is given; what the proper allowance is, has been explained a few lines back. We need not waste time on exemplifying this simple principle; there is so far no real laxity; the writer is simply free.

Laxity comes in when we choose, guided by nothing more authoritative than euphony, to stop an adverbial phrase or adverbial clause, but not to stop it at both ends, though it stands in the middle of its sentence. This is an unmistakable offence against logic, and lays one open to the condemnation of examiners and precisians. But the point we wish to make is that in a very large class of sentences the injury to meaning is so infinitesimal, and the benefit to sound so considerable, that we do well to offend. The class is so large that only one example need be given:

The adverbial phrase is with their triumph over the revolt. But does not belong to it, but to the whole sentence. The writer has no defence whatever as against the logician; nevertheless, his reader will be grateful to him. The familiar intrusion of a comma after initial And and For where there is no intervening clause to justify it, of which we gave examples when we spoke of overstopping, comes probably by false analogy from the unpleasant pause that rigid punctuation has made common in sentences of this type.

Laxity once introduced, however, has to be carefully kept within bounds. It may be first laid down absolutely that when an adverbial clause is to be stopped, but incompletely stopped, the omitted stop must always be the one at the beginning, and never the one at the end. Transgression of this is quite intolerable; we shall give several instances at the end of the section to impress the fact. But it is also true that even the omission of the beginning comma looks more and more slovenly the further we get from the type of our above cited sentence. The quotations immediately following are arranged from the less to the more slovenly.

It will be observed that in the sentence from Mr. Balfour the chief objection to omitting the comma between and and without is that we are taken off on a false scent, it being natural at first to suppose that we are to supply was again; this can only happen when we are in the middle of a sentence, and not at the beginning as in the pattern Cranmer sentence.

The gross negligence or ignorance betrayed by giving the first and omitting the second comma will be convincingly shown by this array of sentences from authors of all degrees.

5. Parenthesis.

In one sense, everything that is adverbial is parenthetic: it can be inserted or removed, that is, without damaging the grammar, though not always without damaging the meaning, of the sentence. But the adverbial parenthesis, when once inserted, forms a part of the sentence; we have sufficiently dealt with the stops it requires in the last section; the use of commas emphasizes its parenthetic character, and is therefore sometimes desirable, sometimes not; no more need be said about it.

Another kind of parenthesis is that whose meaning practically governs the sentence in the middle of which it is nevertheless inserted as an alien element that does not coalesce in grammar with the rest. The type is—But, you will say, Caesar is not an aristocrat. This kind is important for our purpose because of the muddles often made, chiefly by careless punctuation, between the real parenthesis and words that give the same meaning, but are not, like it, grammatically separable. We shall start with an indisputable example of this muddle:

These commas cannot possibly indicate anything but parenthesis; but, if the comma’d words were really a parenthesis, we ought to have would she instead of she would. The four sentences that now follow are all of one pattern. The bad stopping is probably due to this same confusion between the parenthetic and the non-parenthetic. But it is possible that in each the two commas are independent, the first being one of those that are half rhetorical and half caused by false analogy, which have been mentioned as common after initial And and For; and the second being the comma wrongly used, as we have maintained, before substantival that-clauses.

A less familiar form of this mistake, and one not likely to occur except in good writers, since inferior ones seldom attempt the construction that leads to it, is sometimes found when a subordinating conjunction is placed late in its clause, after the object or other member. In the Thackeray sentence, it will be observed that the first comma would be right (1) if them had stood after discovered instead of where it does, (2) if them had been omitted, and any had served as the common object to both verbs.

6. The misplaced comma.

Some authors would seem to have an occasional feeling that here or hereabouts is the place for a comma, just as in handwriting some persons are well content if they get a dot in somewhere within measurable distance of its i. The dot is generally over the right word at any rate, and the comma is seldom more than one word off its true place.

Exactly qualifies and belongs to in so far, &c., not such. The comma should be before it.

For the most part, alone, is the adverbial parenthesis.

But this fault occurs, perhaps nine times out of ten, in combination with the that-clause comma so often mentioned. It may be said, when our instances have been looked into, that in each of them, apart from the that-clause comma, which is recognized by many authorities, there is merely the licence that we have ourselves allowed, omission of the first, without omission of the last, comma of an adverbial parenthesis. But we must point out that Huxley, Green, and Mr. Balfour, man of science, historian, and philosopher, all belong to that dignified class of writers which is supposed to, and in most respects does, insist on full logical stopping; they, in view of their general practice, are not entitled to our slovenly and merely literary licences.

7. Enumeration.

This name, liberally interpreted, is meant to include several more or less distinct questions. They are difficult, and much debated by authorities on punctuation, but are of no great importance. We shall take the liberty of partly leaving them undecided, and partly giving arbitrary opinions; to argue them out would take more space than it is worth while to give. But it is worth while to draw attention to them, so that each writer may be aware that they exist, and at least be consistent with himself. Typical sentences (from Beadnell) are:

Common variants for (a) are (1) Industry, honesty and temperance are essential ... (2) Industry, honesty and temperance, are essential ... (3) Industry, honesty, and temperance are essential.... We unhesitatingly recommend the original and fully stopped form, which should be used irrespective of style, and not be interfered with by rhetorical considerations; it is the only one to which there is never any objection. Of the examples that follow, the first conforms to the correct type, but no serious harm would be done if it did not. The second also conforms; and, if this had followed variant (1) or (2), here indistinguishable, we should have been in danger of supposing that Education and Police were one department instead of two. The third, having no comma after interests, follows variant (3), and, as it happens, with no bad effect on the meaning. All three variants, however, may under different conditions produce ambiguity or worse.

Sometimes enumerations are arranged in pairs; it is then most unpleasant to have the comma after the last pair omitted, as in:

There is a bastard form of enumeration against which warning is seriously needed. It is viewed as, but is not really, a legitimate case of type (a); and a quite unnecessary objection to the repetition of and no doubt supplies the motive. Examples are:

The principle is this: in an enumeration given by means of a comma or commas, the last comma being replaced by or combined with and—our type (a), that is—, there must not be anything that is common to two members (as here, counter-manœuvred with, had, loss) without being common to all. We may say, Moltke had recruited and trained and knew, Moltke had recruited, had trained, and knew, or, Moltke had recruited, trained, and known; but we must not say what the Times says. The third sentence may run, Loss of time and money, and sore trial, or, Loss of time, of money, and of patience; but not as it does.

So much for type (a). Type (b) can be very shortly disposed of. It differs in that the conjunction (and, or, nor, &c.) is expressed every time, instead of being represented except in the last place by a comma. It is logically quite unnecessary, but rhetorically quite allowable, to use commas as well as conjunctions. The only caution needed is that, if commas are used at all, and if the enumeration does not end the sentence, and is not concluded by a stronger stop, a comma must be inserted after the last member as well as after the others. In the type sentence, which contains two enumerations, it would be legitimate to use commas as well as ands with one set and not with the other, if it were desired either to avoid monotony or to give one list special emphasis. The three examples now to be added transgress the rule about the final comma. We arrange them from bad to worse; in the last of them, the apparently needless though not necessarily wrong comma after fall suggests that the writer has really felt a comma to be wanting to the enumeration, but has taken a bad shot with it, as in the examples of section 6 on the misplaced comma.

Of type (c) the characteristic is that we have two or more adjectives attached to a following noun; are there to be commas between the adjectives, or not? The rule usually given is that there should be, unless the last adjective is more intimately connected with the noun, so that the earlier one qualifies, not the noun, but the last adjective and the noun together; it will be noticed that we strictly have no enumeration then at all. This is sometimes useful; and so is the more practical and less theoretic direction to ask whether and could be inserted, and if so use the comma, but not otherwise. These both sound sufficient in the abstract. But that there are doubts left in practice is shown by the type sentence, which Beadnell gives as correct, though either test would rather require the comma. He gives also as correct, Can flattery soothe the dull, cold ear of death?—which is not very clearly distinguishable from the other. Our advice is to use these tests when in doubt, but with a leaning to the omission of the comma. If it happens that a comma of this particular class is the only stop in a sentence, it has a false appearance of dividing the sentence into two parts that is very unpleasant, and may make the reader go through it twice to make sure that all is right—an inconvenience that should by all means be spared him.

Type (d) is one in which the final word or phrase of a sentence has two previous expressions standing in the same grammatical relation to it, but their ending with different prepositions, or the fact that one is to be substituted for the other, or the length of the expressions, or some other cause, obscures this identity of relation. Add to the type sentence the following:

The rules we lay down are: (1) If possible use no stops at all. (2) Never use the second comma and omit the first. (3) Even when the first is necessary, the second may often be dispensed with. (4) Both commas may be necessary if the phrases are long.

We should correct all the examples, including the type: the type under rule (1); the Bryce (which is strictly correct) under rule (3); the Bagehot under rules (2) and (1); and the Balfour under rules (2) and (3); the list two are clearly wrong. The four would then stand as follows:

Learners will be inclined to say: all this is very indefinite; do give us a clear rule that will apply to all cases. Such was the view with which, on a matter of even greater importance than punctuation, Procrustes identified himself; but it brought him to a bad end. The clear rule, Use all logical commas, would give us:

No one would write this who was not suffering from bad hypertrophy of the grammatical conscience. The clear rule, Use no commas in this sort of enumeration, would give:

No one would write this without both commas (after because and why) who was not deeply committed to an anti-comma crusade. Between the two extremes lie cases calling for various treatment; the ruling principle should be freedom within certain limits.

8. The comma between independent sentences.

Among the signs that more particularly betray the uneducated writer is inability to see when a comma is not a sufficient stop. Unfortunately little more can be done than to warn beginners that any serious slip here is much worse than they will probably suppose, and recommend them to observe the practice of good writers.

It is roughly true that grammatically independent sentences should be parted by at least a semicolon; but in the first place there are very large exceptions to this; and secondly, the writer who really knows a grammatically independent sentence when he sees it is hardly in need of instruction; this must be our excuse for entering here into what may be thought too elementary an explanation. Let us take the second point first; it may be of some assistance to remark that a sentence joined to the previous one by a coordinating conjunction is grammatically independent, as well as one not joined to it at all. But the difference between a coordinating and a subordinating conjunction is itself in English rather fine. Every one can see that ‘I will not try; it is dangerous’ is two independent sentences—independent in grammar, though not in thought. But it is a harder saying that ‘I will not try, for it is dangerous’ is also two sentences, while ‘I will not try, because it is dangerous’ is one only. The reason is that for coordinates, and because subordinates; instead of giving lists, which would probably be incomplete, of the two kinds of conjunction, we mention that a subordinating conjunction may be known from the other kind by its being possible to place it and its clause before the previous sentence instead of after, without destroying the sense: we can say ‘Because it is dangerous, I will not try’, but not ‘For it is dangerous, I will not try’. This test cannot always be applied in complicated sentences; simple ones must be constructed for testing the conjunction in question.

Assuming that it is now understood (1) what a subordinating and what a coordinating conjunction is, (2) that a member joined on by no more than a coordinating conjunction is a grammatically independent sentence, or simply a sentence in the proper meaning of the word, and not a subordinate clause, we return to the first point. This was that, though independent sentences are regularly parted by at least a semicolon, there are large exceptions to the rule. These we shall only be able to indicate very loosely. There are three conditions that may favour the reduction of the semicolon to a comma: (1) Those coordinating conjunctions which are most common tend in the order of their commonness to be humble, and to recognize a comma as sufficient for their dignity. The order may perhaps be given as: and, or, but, so, nor, for; conjunctions less common than these should scarcely ever be used with less than a semicolon; and many good writers would refuse to put a mere comma before for. (2) Shortness and lightness of the sentence joined on helps to lessen the need for a heavy stop. (3) Intimate connexion in thought with the preceding sentence has the same effect. Before giving our examples, which are all of undesirable commas, we point out that in the first two there are independent signs of the writers’ being uneducated; and such signs will often be discoverable. It will be clear from what we have said why the others are bad—except perhaps the third; it is particularly disagreeable to have two successive independent sentences tagged on with commas, as those beginning with nor and for are in that example.

It will not be irrelevant to add here, though the point has been touched upon in Understopping, that though a light and-clause may be introduced by no more than a comma, it does not follow that it need not be separated by any stop at all, as in:

9. The semicolon between subordinate members.

Just as the tiro will be safer if he avoids commas before independent sentences, so he will generally be wise not to use a semicolon before a mere subordinate member. We have explained, indeed, that it is sometimes quite legitimate for rhetorical reasons, and is under certain circumstances almost required by proportion. This is when the sentence contains commas doing less important work than the one about which the question arises. But the tiro’s true way out of the difficulty is to simplify his sentences so that they do not need such differentiation. Even skilful writers, as the following two quotations will show, sometimes come to grief over this.

In the first of these the second comma and the semicolon clearly ought to change places. In the second it looks as if Carlyle had thought it dull to have so many commas about; but the remedy was much worse than dullness. Avoidance of what a correspondent supposes to be dull, but what would in fact be natural and right, accounts also for the following piece of vicarious rhetoric; the writer is not nearly so excited, it may be suspected, as his semicolons would make him out. The ordinary sensible man would have (1) used commas, and (2) either omitted the third and fourth denies (reminding us of Zola’s famous j’accuse, not vicarious, and on an adequate occasion), or else inserted an and before the last repetition.

10. The exclamation mark when there is no exclamation.

We must differ altogether from Beadnell’s rule that ‘This point is used to denote any sudden emotion of the mind, whether of joy, grief, surprise, fear, or any other sensation’—at least as it is exemplified in his first instance, given above. The exclamation mark after friend is justifiable, not the other. The stop should be used, with one exception, only after real exclamations. Real exclamations include (1) the words recognized as interjections, as alas, (2) fragmentary expressions that are not complete sentences, as My friend in the example, and (3) complete statements that contain an exclamatory word, as:

The exception mentioned above is this: when the writer wishes to express his own incredulity or other feeling about what is not his own statement, but practically a quotation from some one else, he is at liberty to do it with a mark of exclamation; in the following example, the epitaph-writer expresses either his wonder or his incredulity about what Fame says.

The exclamation mark is a neat and concise sneer at the legal profession.

Outside these narrow limits the exclamation mark must not be used. We shall quote a very instructive saying of Landor’s: ‘I read warily; and whenever I find the writings of a lady, the first thing I do is to cast my eye along her pages, to see whether I am likely to be annoyed by the traps and spring-guns of interjections; and if I happen to espy them I do not leap the paling’. To this we add that when the exclamation mark is used after mere statements it deserves the name, by which it is sometimes called, mark of admiration; we feel that the writer is indeed lost in admiration of his own wit or impressiveness. But this use is mainly confined to lower-class authors; when a grave historian stoops to it, he gives us quite a different sort of shock from what he designed.

11. Confusion between question and exclamation.

We have started with three indisputable instances of the exclamation mark used for the question mark. It is worth notice that the correct stopping for the end of the second quotation (though such accuracy is seldom attempted) would be:—long?’? To have fused two questions into an exclamation is an achievement. But these are mere indefensible blunders, not needing to be thought twice about, such as author and compositor incline to put off each on the other’s shoulders.

The case is not always so clear. In the six sentences lettered for reference, a-d have the wrong stop; in e the stop implied by he exclaims is also wrong; in f, though the stop is right assuming that the form of the sentence is what was really meant, we venture to question this point, as we do also in some of the earlier sentences. Any one who agrees with the details of this summary can save himself the trouble of reading the subsequent discussion.

It will be noticed that in all these sentences except c there is a negative, which puts them, except f, wrong; while in c it is the absence of the negative that makes the question wrong. It will be simplest to start with c. The writer clearly means to let us know that many people see the kingfisher first as a blue streak. He might give this simply so, as a statement. He might (artificially) give it as an exclamation—How many first see him so! Or he might (very artificially) give it as a question—How many do not first see him so?—a ‘rhetorical question’ in which How many interrogative is understood to be equivalent to Few positive. He has rejected the simple statement; vaulting ambition has o’erleapt, and he has ended in a confusion between the two artificial ways of saying the thing, taking the words of the possible exclamation and the stop of the possible question. In a, b, d, and implicitly in e, we have the converse arrangement, or derangement. But as a little more clear thinking is required for them, we point out that the origin of the confusion (though the careless printing of fifty or a hundred years ago no doubt helped to establish it) lies in the identity between the words used for questions and for exclamations. It will be enough to suggest the process that accounts for a; the ambiguity is easily got rid of by inserting a noun with what.

That is the first stage; the resemblance is next increased by inverting subject and verb in the exclamation, which is both natural enough in that kind of sentence, and particularly easy after In that interval. So we get

The words, when the bracketed part of each sentence is left out, are now the same; but the question is of course incapable of giving the required meaning. The writer, seeing this, but deceived by the order of words into thinking the exclamation a question, tries to mend it by inserting not; what ... not, in rhetorical questions, being equivalent to everything. At this stage some writers stick, as Stevenson in f. Others try to make a right out of two wrongs by restoring to the quondam exclamation, which has been wrongly converted with the help of not into a question, the exclamation mark to which it has after conversion no right. Such is the genesis of a, b, d. The proper method, when the simple statement is rejected, as it often reasonably may be, is to use the exclamation, not the Stevensonian question[13], to give the exclamation its right mark, and not to insert the illogical negative.

12. Internal question and exclamation marks.

By this name we do not mean that insertion of a bracketed stop of which we shall nevertheless give one example. That is indeed a confession of weakness and infallible sign of the prentice hand, and further examples will be found in Airs and Graces, miscellaneous; but it is outside grammar, with which these sections are concerned.

What is meant is the ugly stop in the middle of a sentence, unbracketed and undefended by quotation marks, of which examples follow. To novelists, as in the first example, it may be necessary for the purpose of avoiding the nuisance of perpetual quotation marks. But elsewhere it should be got rid of by use of the indirect question or otherwise. Excessive indulgence in direct questions or exclamations where there is no need for them whatever is one of the sensational tendencies of modern newspapers.

Amusing efforts are shown below at escaping the ugliness of the internal question mark. Observe that the third quotation has a worse blunder, since we have here two independent sentences.

13. The unaccountable comma.

We shall now conclude these grammatical sections with a single example of those commas about which it is only possible to say that they are repugnant to grammar. It is as difficult to decide what principle they offend against as what impulse can possibly have dictated them. They are commonest in the least educated writers of all; and, next to these, in the men of science whose overpowering conscientiousness has made the mechanical putting in of commas so habitual that it perhaps becomes with them a sort of reflex action, and does itself at wrong moments without their volition.

It was said in the general remarks at the beginning of this chapter that the systematic use of the colon as one of the series (,), (;), (:), (.), had died out with the decay of formal periods. Many people continue to use it, but few, if we can trust our observation, with any nice regard to its value. Some think it a prettier or more impressive stop than the semicolon, and use it instead of that; some like variety, and use the two indifferently, or resort to one when they are tired of the other. As the abandonment of periodic arrangement really makes the colon useless, it would be well (though of course any one who still writes in formal periods should retain his rights over it) if ordinary writers would give it up altogether except in the special uses, independent of its quantitative value, to which it is being more and more applied by common consent. These are (1) between two sentences that are in clear antithesis, but not connected by an adversative conjunction; (2) introducing a short quotation; (3) introducing a list; (4) introducing a sentence that comes as fulfilment of a promise expressed or implied in the previous sentence; (5) introducing an explanation or proof that is not connected with the previous sentence by for or the like. Examples are:

In the following clear case of antithesis a colon would have been more according to modern usage than the semicolon.

It now only remains, before leaving actual stops for the dash, hyphen, quotation mark, and bracket, to comment on a few stray cases of ambiguity, false scent, and ill-judged stopping. We have not hunted up, and shall not manufacture, any of the patent absurdities that are amusing but unprofitable. The sort of ambiguity that most needs guarding against is that which allows a sleepy reader to take the words wrong when the omission or insertion of a stop would have saved him.

With no comma after unnaturally the first thought is that the agitators not unnaturally consider; second thoughts put it right; but second thoughts should never be expected from a reader.

With no comma after the first word, the sleepy reader is set wondering what simultaneously extensive means, and whether it is journalese for equally extensive.

A comma after together would save us from adding the two sets of years to each other. In the next piece, on the other hand, the uncomfortable comma after gold is apparently meant to warn us quite unnecessarily that here and there belongs to the verb.

The omission of the comma between and and beating would ordinarily be quite legitimate. Here, it puts us off on a false scent, because it allows beating to seem parallel with crossing and object to succeeded in; we have to go back again when we get to the end, and work it out.

The want of a comma between French and demurring makes us assume an absolute construction and expect another subject, of which we are disappointed.

The next two pairs of examples illustrate the effect of mere accidental position on stopping. This is one of the numberless small disturbing elements that make cast-iron rules impossible in punctuation.

That is, a substantival clause out of its place is generally allowed the comma that all but the straitest sect of punctuators would refuse it in its place.

The half adverbs half conjunctions of which therefore and however are instances occupy usually the second place in the sentence. When there, it is of little importance whether they are stopped or not, though we have indicated our preference for no stops. But when it happens that they come later (or earlier), the commas are generally wanted. Therefore in the first of these sentences would be as uncomfortable if stripped as however actually is in the second.

Moved beyond his wont by our English ill-treatment of the dash, Beadnell permits himself a wail as just as it is pathetic.

It is all only too true; these are the realms of Chaos, and the lord of them is Sterne, from whom modern writers of the purely literary kind have so many of their characteristics. Wishing for an example, we merely opened the first volume of Tristram Shandy at a venture, and ‘thus the Anarch old With faltering speech and visage incomposed Answered’:

The modern newspaper writer who overdoes the use of dashes is seldom as incorrect as Sterne, but is perhaps more irritating:

It must be added, however, that Beadnell himself helps to make things worse, by countenancing the strange printer’s superstition that (,—) is beautiful to look upon, and (—,) ugly.

Under these circumstances we shall have to abandon our usual practice of attending only to common mistakes, and deal with the matter a little more systematically. We shall first catalogue, with examples, the chief uses of the dash; next state the debatable questions that arise; and end with the more definite misuses. It will be convenient to number all examples for reference; and, as many or most of the quotations contain some minor violation of what we consider the true principles, these will be corrected in brackets.

1. Chief common uses.

a. Adding to a phrase already used an explanation, example, or preferable substitute.

b. Inviting the reader to pause and collect his forces against the shock of an unexpected word that is to close the sentence. It is generally, but not always, better to abstain from this device; the unexpected, if not drawn attention to, is often more effective because less theatrical.

c. Assuring the reader that what is coming, even if not unexpected, is witty. Writers should be exceedingly sparing of this use; good wine needs no bush.

d. Marking arrival at the principal sentence or the predicate after a subordinate clause or a subject that is long or compound.

e. Resuming after a parenthesis or long phrase, generally with repetition of some previous words in danger of being forgotten.