CHAPTER VII

COMMA, SEMICOLON, COLON, AND PERIOD

THEIR DIFFERENTIATION

As has already been seen, we determine what mark of punctuation is helpful at any point in written or printed language by the sense relation between two groups of words and by the degree of separation between such groups. The degrees of separation cannot be so accurately classified that there will not be much apparent and some real inconsistency in the punctuation of the most careful writer, and much more in the uses of marks by different writers.

We shall consider these points in this chapter, in order to show (1) the cause of this inconsistency; (2) how, in some measure, it may be avoided; and (3) that much of it is wholly immaterial, being not in the least misleading.

A diagrammatic grouping indicated by the coordinate conjunction _and_, with or without one of the four principal marks, may help to illustrate the above points and to differentiate, to some extent, the uses of the marks:

1. ... and ... 2. ... , and ... 3. ... ; and ... 4. ... : and ... 5. ... . And ...

What does the absence of a mark before “and” in Diagram 1, and what does each of the marks in the other diagrams, say to the reader, even before he has seen enough of what follows the conjunction to enable him to apprehend the probable grouping?

We, of course, know that “and” connects words and groups of words which are coördinate in sense, and that such groups are often similar in form. Similarity in form with similarity of relationship in groups of words makes reading easy; and much attention is given to it by good writers.

In Diagram 1 “and” stands between two words or groups of words whose apparent relation is quite unmistakable for their real relation, which was not the case in Sentence 1, which therefore required a mark of punctuation. In other words, no mistake can be made, and therefore no mark is needed. However this diagram may be filled out, “and” joins words or groups of words into a single group, such new group being related to some other word or group in the sentence.

Diagram 2 gives notice of a grouping different from that in Diagram 1; and Diagrams 3, 4, and 5 exhibit still other groupings.

We saw, particularly in Sentences 7, 7-1, and 7-2, a simple grouping which plainly required another and a larger mark than the comma. We saw another kind of grouping in other sentences which also unmistakably required a mark other than the semicolon.

We now come to discuss the basis of differentiating our principal marks, sometimes at points where the choice of a mark is not always unmistakable. In some cases the choice is based upon well-defined sense relations; in other cases it is based upon shades of meaning or upon degrees of separation not sufficiently sharp to make the choice easy, especially under _rules_ for punctuation.

Sentence 1-1 is a good illustration of Diagram 2; and yet the apparent relation between “children” and their “mothers” might so strongly impress itself upon the mind of an inexperienced reader, especially if reading the sentence at sight and aloud, that the comma would not prevent a momentary wrong combination of these words. A semicolon before the conjunction would prevent the liability to such error, and probably do so even with the most inexperienced reader. Thus we have a choice between the comma and the semicolon in Sentence 1; and this choice rests upon the degree of liability to error, which cannot be the same with different classes of readers. Liability to error in grouping is increased by a closeness of _apparent_ relations that are _wrong_ relations, and also by a reader’s lack of expertness in grouping language at sight. In view of these facts, we think the use of a semicolon in Sentence 1 better punctuation than the use of a comma; but the use of the semicolon in this particular sentence should not be made to justify a rule requiring a semicolon between every two clauses connected by a conjunction, or between two clauses so connected, one of which contains one or more commas.

Our next sentence is taken from an essay, in a high-class magazine, by the professor of English in a leading eastern university; and this essay has for its title “Writing English”:

58. The boy must be able to say what he knows, or write what he knows, or he does not know it.

As written and punctuated, this sentence contains three apparently coördinate groups of words, the groups being connected by “or,” and the grouping indicated by commas. It is, however, composed of only two main groups, the first of the two groups being regrouped into other groups. Such grouping should always be shown by the punctuation; for, when it is not so shown, the sentence cannot be read without some distractive effect upon the reader and, if read aloud, also upon the hearer.

The failure of this sentence to read smoothly, or without distractive effect, arises from the usual cause: the form of the language leads the reader to look forward to a mode of development of the sentence which does not follow,—that is, the _apparent_ mode of development, as exhibited in the grouping, is not the _real_ mode of development. As each of the second and third groups is introduced by “or” and preceded by a comma, the reader is led to look for the same relation between the second and third groups that exists between the first and second. He is thus led to expect a group in a series, which does not follow.

The distractive effect of the sentence can be removed in some degree by writing it in the following form, the added “to” giving the second group a form similar to that of the first:

58-1. The boy must be able to say or _to_ write what he knows, or he does not know it.

In this condensed form the sentence lacks, at the point of condensation, the emphasis of the original; and therefore the sentence is not so effective in its statement of a truth. If, then, the original form is preferred, its distractive effect can be removed by the use of punctuation that properly groups its parts:

58-2. The boy must be able to say what he knows, or write what he knows; or he does not know it.

The semicolon here _disconnects_ the second group from the third, giving the third group connection with a larger group, the first and second combined.

The need of the semicolon in No. 58-2 is not determined by the presence of the preceding comma, but mainly by the apparent relations of the groups, such apparent relations being emphasized by the presence of the second “or,” which makes the three groups appear to be in a series. If these groups were in a series, the second “or” would connect the second and third groups. The semicolon at once throws the grouping into two parts, thus making the second “or” connect all that precedes with all that follows.

Although the writer of Sentence 58 appears to have placed no value upon the _like_ formation of the first two groups of this sentence, nor to have recognized the differentiation of the comma from the semicolon, he uses both of these devices in a sentence, quite like the above, in the next paragraph of his essay. It is as follows:

59. It is here that composition is of service to the imagination, and incidentally to culture; and I should speak more largely of this service if there were space in this essay to bring forward all the aspects of college composition.

The need of the semicolon in both Nos. 58 and 59 is, we think, unmistakable.

The comma in No. 59 has no value for the reader. There is no _apparent_ and _wrong_ meaning here requiring this comma; and it is difficult to see how any wrong grouping, in the absence of the comma, is likely to be made by the reader. If this comma were challenged for its meaning we do not know what answer it could give.

The absence of a comma after “imagination” would give notice to the reader that what has been said about “imagination” is to be said about something else in the _and_ relation to “imagination.” The writer, wishing to qualify the added statement, does so before the statement is made, thus throwing matter (one word in this case) between what has preceded and what is to be grouped with “imagination,” as shown by the absence of a mark before “and.” This makes the qualifying word a modified parenthesis. Such treatment of the word, indicated by marks, gives a shade of meaning difficult to express in language. The difference is, perhaps, in the emphasis gained by the punctuation.

We should write the sentence in the following way:

59-1. It is here that composition is of service to the imagination and, incidentally, to culture; and I should speak more largely....

Our next sentence, which is taken from a well-known literary paper, is punctuated so badly that we wonder how it could pass an experienced proof-reader:

60. Providence made him a waif on London streets and later, a waif on the ocean, she taught him the feeling of a rope’s end on a naked back.

When the reader reaches “and” in this sentence he needs to be informed by a mark of punctuation that what immediately follows the conjunction looks forward, not backward, for its connection. Such notice is always imperative when the words immediately following the conjunction may readily be attached by the reader to what precedes, but belong to what follows. In this sentence “a waif on the ocean” readily becomes the object of “made”; and even a very careful reader might so connect it, not discovering his mistake until he reached “she taught him.” “Later” apparently looks backward to “made”; but its connection is forward.

Properly punctuated the sentence reads as follows:

60-1. Providence made him a waif on London streets; and, later, a waif on the ocean, she taught him the feeling of a rope’s end on a naked back.

“Later” takes a comma before it because it is out of its natural order and also because it is placed in a position where it obstructs the smoothness of the language. The position of both “later” and “a waif on the ocean” gives force and beauty to the language; but they entangle, to a certain extent, the principal words of the sentence, thus requiring marks of punctuation to point out the proper relation between such words, as in Sentence 4-2.

Diagram 4 covers the punctuation exemplified in Sentence 6,—that is, punctuation based upon the rank of the colon.

The use of the colon that is properly differentiated from the uses of the semicolon and the period, respectively, is the use determined by its rank. This particular use is exemplified in Sentences 6 and 6-2 and, to some extent, in Sentences 26-3 and 27-1. Writers who ignore the latter use of the colon—and they are many in number—apparently ignore the real relations between the parts of a sentence or of a composition; or, if they themselves see and appreciate such relations, they are willing to leave them obscurely expressed. It is quite probable that many writers do not know that marks of punctuation have inherent meanings which clearly express these relations. The lack of such knowledge is explained by the almost total absence of any discussion of the matter in the text-books on punctuation.

Our statement that Sentences 26-3 and 27-1 illustrate only “to some extent” the rank of the colon among the principal marks, may need explanation. In Sentences 6 and 6-2 the colon is used because a mark larger than the semicolon is required to group the parts of each sentence, one part being subdivided by semicolons. This use is wholly one of rank, and clearly differentiates the use of the mark. It is another sense relation of the parts of each of the other sentences (Nos. 26-3 and 27-1) that requires the colon, and not the presence of a semicolon in one of the parts. The colon would still be the only mark with an inherent meaning expressive of the relation between the parts even if there were no other mark in either sentence. The colon would still be the proper mark in No. 26-3 if it ended with “old,” and in No. 27-1 if it ended with “evasions.” The relation between the two groups following the colon in No. 26-3 is the _and_ relation, each group being an expansion of a picture in the main group,—that is, the first group (it never grows old) explains “perpetual charm,” and the second group (it never loses its novelty) explains “singular charm.” In No. 27-1 a conjunction (but) is present to show the sense relation at that point.

Our next sentence, which is from Balzac, will illustrate the points we are considering, and also show the difficulties the punctuator meets in the punctuation of language which he may not change:

61. He was all that a hero should be; he suffered long, and came into his own, and he has been, in a measure, the world’s ideal of a heroic man of arms.

According to most rules of punctuation based upon grammatical relations, this punctuation is correct; according to the principles of punctuation based upon sense relations, this punctuation is very bad. The obscurity in the meaning of the language cannot be wholly removed by punctuation. Two simple statements are made in the sentence, and are coördinated by the second conjunction (and); but the language is so grouped by the punctuation as to conceal this fact. We can show the meaning by a different mode of punctuation, which, however, is not good punctuation. Such punctuation will show the relation because the meaning of the marks used is familiar to most readers:

61-1. He was all that a hero should be (he suffered long, and came into his own); and he has been, in a measure, the world’s idea of a heroic man of arms.

The parentheses show the relation of the matter so enclosed to what has preceded; but the matter so set off is not a parenthesis, and therefore the use of the marks of parenthesis is not good punctuation. It is true that the matter so set off is explanatory matter; but it is an expansion of the idea preceding it, and therefore should be preceded by a colon.

Having determined that a colon is the proper mark after this first group of words, we must then determine where the second group ends. It is quite evident that “came into his own” is joined to “suffered long,” even though _coming into one’s own_ is no part of what constitutes a hero. We simply accept the words at Balzac’s estimate.

The meaning of the next group, introduced by the second “and,” is unmistakable; and the group is coordinate in sense, and is connected by the conjunction, with all that precedes. If the group preceding the conjunction is subdivided by a colon, even though the colon here is not used solely because of its rank, it is well to ignore this technicality, and use a period, thus unmistakably and properly grouping the language.

This reasoning would give us the following punctuation:

61-2. He was all that a hero should be: he suffered long, and came into his own. And he has been, in a measure, the world’s ideal of a heroic man of arms.

The period here connects the second sentence with _all_ of the preceding sentence, at once disconnecting the second sentence from the second clause of the first sentence, of which it is made an unmistakable part in the original sentence (No. 61).

This sentence illustrates the loose grouping common to many of our best writers; and it also shows that marks of punctuation, if possessing clearly understood meanings, cannot be adjusted to loose groupings.

If the colon in No. 61-2 were considered like the colon in No. 25, and thus held subordinate to the semicolon, a semicolon would be preferable to the period that follows; yet very few readers would grasp such fine use of a mark in a sentence like No. 61-2.

Our next sentence illustrates a fault that is practically the opposite of the above; and its punctuation is much more common and even more distracting. It is the fault of splitting up a sentence, making two sentences out of one, or of carrying a part of one sentence over into another sentence, making two incomplete sentences. This fault grows out of an effort to use short sentences, which some writers think essential to clearness. It is quite safe to say few sentences are too long if the relations between their parts can clearly be shown by marks of punctuation.

We take the sentence from _The Outlook_. It appears in Dr. Abbott’s “Reminiscences”; and, as Dr. Abbott is noted as an unusually clear and forceful writer, the sentence, with its context, is of special interest:

62. There were lawyers [in 1853-59] who promoted quarrels to get fees. But they were the pariahs of the profession. The best lawyers were peacemakers, and though, of necessity, professional partisans when engaged in litigation, they were generally honorable partisans.

That the third of the above sentences grows out of the first is clearly shown by the contrast between the kinds of lawyers mentioned: the former promoted _quarrels_; the latter were _peacemakers_. Note also that each class is qualified: the first is qualified by “but they were the pariahs”; the second by “and were generally honorable partisans.” The “and,” which introduces a really qualifying clause, might properly have been “but,” to show that as “peacemakers” these lawyers surrendered no rights of their clients, _but_ contended for such rights when necessary to become “professional partisans.”

As the sentences are written, the third sentence does not grow out of the sentence (the second) preceding it, but out of the first sentence; and such remoteness of parts standing in the close relation required in a contrast, is distracting.

We used a period in No. 61-2 because “and” there connects two complete sentences standing in the _and_ relation to each other. We object to the use of a period in No. 62 where the conjunction (but) connects two sentences which are properly only one, and should be so written.

The points we wish to make may be seen by a slight change in the sentence:

62-1. There were lawyers who promoted quarrels to get fees; but they were the pariahs of the profession. The best lawyers were peacemakers; but, when engaged in litigation and it became necessary, they were professional partisans, though they were generally honorable partisans.

In order to see more clearly the differentiation between relations that require the semicolon and those that require the colon, let us study these relations in two sentences properly punctuated with each of the two marks:

63. The only true equalizers in the world are books; the only treasure-house open to all comers is a library; the only wealth which will not decay is knowledge.

64. Here was a superintendent worth having: when he did not find good tools, he made them.

Why not use commas or periods in No. 63? Why not use a semicolon in No. 64?

Perhaps we have not made, and cannot make, the differentiation between the comma and the semicolon sufficiently definite to answer the first question, especially as the formation of each of the three clauses is such that the liability to error in reading them is very slight. It would be rather dogmatic to assert that the use of commas here would be poor punctuation; yet the degree of separation between the groups is so marked as to make the relations more like the relations between groups that we have seen to require a semicolon than like relations between groups separated by a comma. The degree of separation here almost approaches the degree that takes the period; and therefore the semicolon seems much better than the comma.

But why not periods? Two positive answers may be given. The three clauses constitute one complete idea, and therefore make a sentence. If set off by periods, thus making three sentences, no fourth sentence could be formed to follow the third and maintain sentence unity; for the fourth sentence would necessarily grow out of the third, while its logical relation would be to the three clauses. Such a fault would be that seen in Sentence 62.

Why not a semicolon in No. 64? Because, under our classifications and definitions, the colon, when its use is not determined by its rank, as it is in Sentence 6, is used to show a relation that is practically that of apposition, exhibited, in different forms, in Sentences 25 and 26-3. “A good superintendent” is explained by the expanded clause following the colon, just as “articles” in No. 25 is explained by the details that follow the colon.

In most sentences like No. 64 the second clause can be tied to the first by a conjunction (for, as, etc.); but the exact idea is not thus expressed. While this would sometimes show the sense relation between the clauses, it would distinctly weaken the force and beauty of the language, tending to make the second clause a specific, rather than a general, statement.

The differentiation between the semicolon and the colon in Nos. 63 and 64 is very clear if we recognize the colon as the mark of apposition or explanation. In No. 63 other clauses are added to the first to add other thoughts; in No. 64 the added clause repeats the thought of the first by way of amplification of such thought.

The greatest value of a mark of punctuation lies in its ability to indicate a meaning that depends upon the relation of one group of words to another; and so long as language has a literary value, just so long will the colon be useful as an aid in the expression of thought. This is illustrated in our next sentence (No. 65), where a semicolon, instead of the colon, would suggest the _and_ relation, and thus make havoc of the sense:

65. Ethics has no summons to righteousness: it knows nothing of truth spiritually discerned.

If a period were used here, instead of the colon, the relation between the two sentences (clauses formed into sentences by the use of the period) might easily be taken for the _and_ relation. With such relation between the clauses, the first clause would make a vague assertion, which, without an explanation, would convey no definite meaning.

In our discussion of Sentence 6 we saw that the colon, preceding “but,” made the _but_ relation extend back over several clauses separated by semicolons. “But” may begin a sentence, a paragraph, a part of a book, or even a volume. When beginning any one of such groups, the _but_ relation should be a well-defined sense relation between the groups so connected. If “but” begins a paragraph, the sense relation must be between the two paragraphs connected by the conjunction. Such relation is rarely sustained between long paragraphs; and it does not exist between sentences nearly so often as it is indicated by the use of the conjunction and the period.

_But_, _and_, and _for_ are the principal conjunctions thus used, although others are not infrequently so used.

COMMA AND SEMICOLON

The differentiation of the semicolon from the colon, and the colon from the period, seems to be so well marked that a choice between them is rarely difficult to make in the punctuation of language with proper sense relations between its parts; but the differentiation between the comma and the semicolon is not so well defined, yet it is generally very clear in well-written English. Here the choice of a mark is determined, very largely, by the degree of separation between the groups. In order further to illustrate this point, let us compare No. 63 with a sentence which is quite like it, and yet requires only a comma between its clauses:

66. Experience is fallacious, and judgment difficult.

The first three words of the above form a clause which so completely expresses a thought that the addition, by means of the “and,” of another adjective to follow “fallacious,” is hardly suggested. On the other hand, this degree of completeness might suggest that what is to follow will be a complete clause so changing the direction of the thought as to require a semicolon to show the fact of the change. Let us look at the sentence diagrammatically:

66-1. Experience is fallacious and....

How will this sentence probably continue? This question confronts every intelligent reader, whether he realizes it or not, when reading the complete sentence. If the mark before “and,” or the absence of a mark, conveys to him no information upon this point, he learns only through a tiresome mental process of ascertaining the correct relation of the words that follow “and” to the words that precede it. This process is one of holding in suspense two or more possible combinations. It is always distractive, and very often ends in a wrong combination. A semicolon before “and” would suggest that the preceding group of words is complete in itself, and is to be followed by a complete and coördinate thought (“and” indicates the coordination) in the proper sense relation with the group that has preceded. It would at once preclude the expectation of an adjective coördinate with “fallacious.”

We may complete the sentence thus:

66-2. Experience is fallacious; and therefore it is difficult for us to form correct judgments based upon experience.

To exhibit at once the sense relation between the two groups of words in the above (the relation of correct judgment to experience), we inserted the word “therefore”; to show that “and” is to be followed by a group grammatically coördinate with what precedes it, and of like form, we used a semicolon.

If the relation between the groups were closer than the relation suggested by the semicolon, but not so close as indicated by the absence of a mark, the comma would logically be the proper mark. It would suggest neither a word to be grouped with “fallacious” nor a clause to be grouped with all that precedes “and.” It would suggest a different development of the sentence. Such development would tie what follows to a _part_, not to the _whole_, of what precedes; and, as there are but three words in this sentence preceding the mark, what follows must be tied to one of them, and not to the whole, as in No. 66-2. Two sentences will serve to illustrate the point:

66-3. Experience is fallacious, and may not safely be depended upon in the formation of correct judgment.

66-4. Experience is fallacious, and unreliable as a basis of correct judgment.

In No. 66-3 the comma notifies the reader that another adjective in the _and_ relation to “fallacious,” is not to follow.

In No. 66-4 the comma is used to confine “as a basis of correct judgment” to “unreliable,” thus disconnecting it from “fallacious.” This makes the language say that experience is fallacious _in all things_, and unreliable _only_ as a basis of correct judgment.

If the reasoning in our discussion on the differentiation of the comma from the semicolon, be correct, it seems to establish a rule that the semicolon is the mark of choice between two clauses, whether joined by a conjunction or not. It is not entirely the length or the character of the groups of words that determines the choice. The liability to make wrong combinations, especially such combinations as are suggested by the apparent meaning of the language, is to be lessened by the mark.

In the absence of a mark in Sentence 1, a momentary wrong combination is quite unavoidable, because perfectly good sense is made by the wrong grouping. As a comma before “and” may not give to every reader notice of the development of the sentence, and as a semicolon could hardly fail to do so, it would seem to be the better mark.

In neither No. 66-3 nor No. 66-4 can the reader definitely interpret the meaning of the comma until one or two words following the conjunction are reached. As the process of determining a relation from a mark often requires the reader to consider the mark and one or two words following it, which process becomes almost instantaneous, a comma may thus convey its purpose as readily as the semicolon in No. 66-2 tells its purpose.

A slight change in the language of Sentence 1 will not change either the names of the two groups (clauses) or their relation (coordination); but it will remove the liability, even the possibility, of such wrong combination as is suggested in Sentence 1:

66-5. Respect the rights of children and you will gain their respect.

In this sentence, a mark before “and” is not really needed as a warning to the reader not to connect “you” with “children,” thus making “you” an object of the preposition “of.”

The similarity in form between Nos. 1 and 66-5, with a mark of punctuation imperative in the first, suggests the use of a mark in the second. As the comma is here quite sufficient to prevent a wrong combination by the reader, the comma naturally becomes the mark of choice; and good convention confirms this choice, while convention not so good seems to ignore the need of the semicolon in similarly formed sentences in which the _apparent_ and _wrong_ grouping is much more marked than in Sentence 1.

Sentences 66 and 66-5 stand at one extreme of the class of sentences in the punctuation of which we seek to differentiate between the comma and the semicolon. The relation is made quite unmistakable by the shortness and completeness of the first group of words, and also by the fact that the word in each sentence following the conjunction does not suggest, even in the slightest degree, a connection for itself with anything that precedes.

Let us consider a sentence at the other extreme, where a wrong combination is wholly unavoidable without a semicolon, and quite suggestive with one:

66-6. I do not mean that the people are conscious of this fact; but that the leaders of the people are conscious of it, I think, there is no doubt.

In this sentence a semicolon is quite indispensable in order to disconnect the second group beginning with “that” from “mean,” to which the preceding similarly formed group, beginning with “that,” belongs.

Between these extremes are many sentences which may take either the comma or the semicolon without distractive effect. The close punctuator will generally prefer the semicolon; the open punctuator, the comma. (For a discussion of close and open punctuation, see Chapter XV.)

The degrees of variation in the relationship between parts of language are so great that a differentiation between the comma and the semicolon is at times almost impossible. Fortunately, a quite indiscriminate use of these two marks in the class of sentences under consideration is not always misleading or distractive to the reader; but the indiscriminate use of marks tends to lessen the importance attached by a reader to punctuation.

EXAMPLES

1. He suffered much, and he also suffered long.

2. Virtue is intolerant of vice, and virtue is just as contagious as vice ever was.

3. Be the first to say what is self-evident, and you are immortal.

4. In some states the legislatures meet annually, and in others biennially.

4-1. In some states the legislatures meet annually; in others, biennially.

5. Want of intellect makes a village an Eden, a college a sty.

In each of the first four examples above the relation between its groups is indicated by a conjunction; and the grouping is so readily apparent that the comma serves to show it.

In No. 4-1, in the absence of a conjunction, the semicolon at once shows that another clause is to follow as in No. 7, below. Such clause, however, is contracted by the omission of a group of words common to the two clauses. This omission is indicated by a comma in the second clause, which also shows that the two words in the group are not in any grammatical relation to each other.

In the first group of No. 5 we have a double object of the verb (explained later as requiring no comma), and the object is repeated, in form, in what follows the comma. The comma may be said to be used to indicate the omitted verb, or, as in Sentence 1-1, to indicate the relation.

6. The sentences of Seneca are stimulating to the intellect; the sentences of Epictetus are fortifying to the character; the sentences of Marcus Aurelius find their way to the soul.

7. In the world of reality suffering is not a thing to be read or heard or talked about, but a living truth. Being defied, it maketh for bitterness; or ignored, for selfishness; or accepted, for wisdom.

8. Venice, it has been said, differs from all other cities in the sentiment which she inspires. The rest may have admirers; she only, a famous fair one, counts lovers in her train.

9. Fame is what you have taken, Character’s what you give; When to this truth you waken, Then you begin to live.

10. Homer has not Shakespeare’s variations: Homer always composes as Shakespeare composes at his best; Homer is always simple and intelligible, as Shakespeare is often; Homer is never quaint and antiquated, as Shakespeare is sometimes.

11. There are many beautiful letters in Cary’s “Life of Curtis”; there is no other so beautiful as that written just after the death of Lincoln, nor is it possible to read it without a great trembling of the heart.

12. If there is ever a time to be ambitious, it is not when ambition is easy, but when it is hard. Fight in darkness; fight when you are down; die hard, and you won’t die at all.

The comma before “and” in No. 12 divides the _semicolon_ group into two parts, and does not stand between groups coordinate with the first and second groups.

“Die hard, and you won’t die at all” is really a bull; and the incongruity of ideas might well be expressed by a dash before “and.”

13. All association [among people] must be a compromise; and, what is worse, the very flower and aroma of the flower of each of the beautiful natures disappears as they approach each other.

14. All religions, even the most conservative and traditional, are in constant flux, they either advance or decay.

The above sentence is thus punctuated in the Enclycopedia Britannica and the New Standard Dictionary. We know of no meaning of the comma that indicates the real sense relation between “in constant flux” and what follows. Nor is this relation that shown by the colon; for a thing,—for instance, the ocean,—may be _in constant flux_, and yet not either _advance_ or _decay_.

The meaning of the language requires “and” after “flux,” preferably with a semicolon before it.

15. Lily Dyer was a favorite with the village folk; she had just the qualities to arouse admiration. She was good and handsome and smart.

In the above example the _semicolon_ relation does not exist between the groups of words (clauses) between which it is placed, for the second group does not add something to the first to make a complete thought. Moreover, the third group bears an unmistakable relation to the second group; and such relation is not the _period_ relation. The language needs regrouping, to show the sense relations:

15-1. Lily Dyer was a favorite with the village folk. She had just the qualities to arouse admiration: she was good and handsome and smart.

16. It is a party which is as yet without a name. And what is even stranger, without a nickname.

No. 16 is a somewhat extreme, though common, result of seeking to make one’s sentences short. No mark at all is needed, or permissible, before “and” in such a sentence:

16-1. It is a party without a name and, what is even stranger, without a nickname.

17. Polemic is always dreary. Devotion always interests. This is because men are nearly always wrong when they want their own way, and always partly right when they worship.

As the third sentence in the above example does not grow out of the sentence preceding it, sentence unity is destroyed; and, besides, the use of “this” clearly shows reference to a thought expressed in the _two_ preceding sentences. A semicolon is the proper mark after “dreary,” to make one sentence of the two, out of which the third grows.

18. Milton does not paint a finished picture, or play for a mere passive listener. He sketches, and leaves others to fill up the outline. He strikes the key-note, and expects his hearers to make out the melody.

In this example sentence unity is wholly disregarded: the first picture is put into one sentence, instead of two; the second picture is put into two sentences, instead of one. Moreover, as the second picture is but an amplification of the first, both pictures may be put into one sentence, and should be so put, in accordance with the punctuation we have discussed.

As now punctuated, the second sentence does not grow out of the first, but out of the first half of the first; and the third sentence, instead of growing out of the second, grows out of the second half of the first, ignoring the second sentence entirely.

Uniting the second and third sentences by a semicolon will improve the grouping and the punctuation; but making one sentence out of the three is still better:

18-1. Milton does not paint a finished picture, or play for a mere passive listener: he sketches, and leaves others to fill up the outline; he strikes the key-note, and expects his hearers to make out the melody.

The relations between the groups may be shown in another way; but such mode of expressing thought is not tolerable:

18-2. Milton does not paint a finished picture (he sketches, and leaves others to fill up the outline), or play for a mere passive listener (he strikes the key-note, and expects his hearers to make out the melody).