English spelling and spelling reform

by Thomas Raynesford Lounsbury

FOOTNOTES:

[15] Swift. Journal to Stella, December 14, 1710.

CHAPTER III
THE ORTHOGRAPHIC SITUATION

I
THE PROBLEM

It is with a good deal of hesitation that I approach this part of my subject. To treat it fully, to consider it in all its details, would require a familiarity with the history of sounds, with their precise values, and with the proper way of representing these values, to which I can lay no claim. Though I have given some time to the study of this branch of the general question, I am well aware that my knowledge of it is not the knowledge of a professional, but of an amateur. It is only when I read the attempts of the assailants of spelling reform to write what they are pleased to call phonetically, that my own slender acquaintance with this field of research looms up momentarily before my eyes as endowed with colossal proportions. Fortunately, intimate familiarity with this particular part of the subject is not needed for the end had here in view. To point out the evils afflicting our present system is possible for him who is unable to prescribe a remedy. This is the special task which I set before myself in this section.

It is essential, in the first place, to have clearly before our minds the nature of the problem with which we are called upon to deal. The general statement about it may be and often is summarized in a few words. We are told that in English the same sound is represented by half a dozen signs, and the same sign is used to denote half a dozen sounds. This is all true. Unfortunately, to the vast majority of men it conveys no definite idea. It certainly would not bring clearly before them much conception of the real difficulty. Some of them would even be puzzled to explain what is meant here by the word sign. Most of them have no knowledge whatever of the number and quality of the sounds they use. This remark is not intended as a reproach. Their condition of ignorance is due to no fault of their own. The existing orthography does not content itself with hiding from the ordinary eye all knowledge of phonetic law; it puts a stumbling-block in the way of its acquisition. Accordingly, it does more than permit ignorance of the subject; it fosters it. Men are not led to consider even the most aggressively prominent facts of their own utterance. I have known intelligent young persons, of much more than ordinary ability, who had never learned as a matter of knowledge that the digraph th has two distinct sounds; that were thin pronounced as is then, or then as is thin, we should have in each case another word than the one we actually possess. They had never confused the two in their usage, but as little had they been in the habit of remarking the difference between them. Consequently, when it was brought directly to their attention, it came upon them as a sort of surprise. If a distinction which lies on the very surface could so easily escape notice, what hope can be entertained of gaining a realizing sense of those subtler ones which abound on every side?

The first point, therefore, to be made emphatic is that there is a large number of sounds in the speech and but a limited number of signs in the alphabet. The number of sounds has been variously estimated. It depends a good deal upon the extent to which the orthoepic investigator is disposed to recognize differences more or less subtle and the weight he assigns to each. In general, it may be said that usually the lowest number given is thirty-eight, and the highest forty-four. A very common estimate puts them at forty-two. Exactness on this point is not necessary for the purpose here aimed at, and for the sake of convenience the whole number of sounds will be temporarily assumed to be forty, more or less. To represent these forty sounds we have nominally twenty-six letters. Really we have but twenty-three. Either c or k is supernumerary, as are also x and q.

Here, then, lies the initial difficulty. The Roman alphabet we have adopted has not a sufficient number of letters to do the duty required of it. For us its inability has been further aggravated by the loss of two signs which the language had originally, or acquired early in its history. For the disappearance of these there was later in another quarter a partial compensation in the differentiation of i and j, and of u and v. Of the two vanished signs one was a Rune, called “thorn” or the “thorn letter,” þ, the other a crossed d, represented by ð. They were or could be used to represent the surd or hard initial sound heard in thin just mentioned, and the corresponding sonant or soft sound heard in then. These two letters, unknown to the Roman alphabet, were allowed to die out of general use in the fifteenth century. Of both the digraph th took the place. Yet in one way the so-called thorn letter has left behind a memorial of itself. Its form had something of a resemblance to the black-letter character y. Consequently, when “thorn” ceased to be used, y was at times substituted for it. Especially was this true in the case of the words the and that. These were frequently printed as yᵉ and yᵗ. This form of the latter word disappeared after a while, not merely from use, but practically from remembrance. Ye, however, in the sense of the, but with its initial letter given the sound of y, is fondly cherished and sometimes employed by certain persons, who indulge in the delusion that by so doing they are writing and talking Old English.

The use of this one digraph to represent these two distinct sounds inevitably tends to create uncertainty of pronunciation, if not to produce confusion. We can see this fact exemplified in the case of such words as tithe and path and oath and mouth. In these th has in the singular the surd sound, in the plural the sonant. The proper way of pronouncing them has therefore to be learned carefully in each individual instance, for there is nothing in the spelling to indicate it. Usage in truth is very fluctuating with respect to some of the words in which this digraph appears. In consequence, the question of their pronunciation begets at times much controversy. Still, compared with the uncertainty attending other signs, the perplexities caused by this are of slight importance.

The lack of a sufficient number of signs to indicate the sounds is therefore the first difficulty to be encountered. But this is a defect which English shares with several tongues which have adopted the Roman alphabet. There is another characteristic which belongs to our language exclusively. This is the progressive movement which has gone on in the case of some of the vowel-sounds. In the historic development of English pronunciation several of these have lost their original values. This has caused them not merely to deviate from the sounds they once had in our own speech, but has also brought them out of harmony with those of the cultivated tongues of modern Europe. In none of these have the original values experienced any such disturbance. Such a condition of things is so peculiar to our language, it complicates the whole orthographic situation so thoroughly, that it demands first consideration in any discussion of the various problems that need to be solved. Let us give briefly, then, the most important facts in regard to the changes which have taken place in the history of these sounds.

II
MOVEMENT OF VOWEL-SOUNDS

The first vowel-sound of the alphabet—the a heard in father and far—has been aptly styled “the fundamental vowel-tone of the human voice.” But the noticeable fact about it in English is that it has not only gone largely out of use already, but that it tends to go out of use more and more. Once the most common of articulate utterances, it has now become one of the rarest. In reducing the employment of it English has gone beyond all other modern cultivated tongues. The decline in its use has been steady. “In the Sanskrit,” says Whitney, “in its long and short forms it makes over seventy per cent. of the vowels and about thirty per cent. of the whole alphabet.” In examining his own utterance he rated the frequency of its occurrence at a little more than half of one per cent. This may be taken as fairly representative of the fortunes which have generally befallen the sound. Different parts of the English-speaking world preserve it, indeed, in different degrees. In Great Britain—if I can take as typical of all persons the pronunciation of it furnished to my own ears by a few—it is retained more fully than in the United States. But even there it has for a long period been disappearing. There is no reason to suppose that with our present orthography this process will not continue to go on.

The loss of this sound would assuredly be a great calamity to the speech. The coming of that day may be distant; it is to be hoped that it will never come at all. Yet owing to the incapacity of our orthography to represent pronunciation strictly, and therefore hold it fast permanently, the sound is certainly in danger of following to the very end the road on which it has long been travelling. It shows every sign of steady though slow disappearance. Once it was heard generally in many classes of words where it is now never heard at all. Such, for instance, was the case when a was followed by n, as in answer, chance, dance, plant; by f, as in after; by s, as in grass, glass, pass; by st, as in last and vast. More than a century ago the lexicographer Walker contended that this sound must formerly have been always heard in these and such like words, because it was “still the sound given to them by the vulgar, who are generally the last to alter the common pronunciation.” There can be little doubt of the fact. In truth, Doctor Johnson distinctly specified rather, fancy, congratulate, glass among others as having it. Walker added that “the short a in these words”—those mentioned above—“is now the general pronunciation of the polite and learned world.” Hence, he felt justified in asserting that the ancient sound “borders very closely on vulgarity.”

This same result is showing itself in the instances where the vowel is followed by other letters or combinations of letters. Before lf and th—which can be illustrated respectively by calf, half, and by path, bath—the original sound, once generally heard, has given way largely and is still giving way. There are certainly many parts of the English-speaking world where the older pronunciation of it would be the exception and not the rule. The most effective agent in retaining it is a following r. In this case the sound is heard in no small number of words, as may be seen, for illustration, in bar and car. Another agency working for its retention, though far less powerful than the preceding, is a following l, as in balm and calm. But in this second case the sound is even now threatened with extinction. It exhibits in many places weakness of hold upon the utterance. Hence, it may come to take the road already trodden by other words in which it once showed itself. In the case of some of these, as a result of diminishing use, the sound, when heard, for illustration, in words like half and calf, is already looked upon by many as an affectation. Should such a feeling about it come not only to exist but to prevail when the vowel is followed by lm, its doom would be sealed. To hear psalm pronounced as the proper name Sam is still hateful to the orthoepically pure. Such a usage can as yet be politely termed a provincialism, or, insultingly, a vulgarism. Yet against the levelling tendency of an orthography which does not protect pronunciation, it is possible that the earlier sound of a in these words may not be able to hold out forever.

So much for the first vowel of the alphabet. We are as badly off, though in a different way, when we come to the second. It emphasizes the degeneracy which has overtaken our whole orthographic and orthoepic system that the name we now give to the first vowel was originally and still is scientifically the long sound of the second. The respective short and long values of this are heard in the words met and mate. In them are indicated the two sounds which the second vowel once had with us, and which it still retains in other cultivated tongues. The short sound continues to exist in all its primitive vigor, but the long sound is now very generally denoted by a. E itself no longer has it, save in the exclamation eh, and in certain cases where it is followed by i or y, such as vein and rein, and they or obey. Perhaps, indeed, it would be better to say that, strictly speaking, this letter by itself never indicates the sound at all; for the digraphs ei and ey, as we shall see later, have various distinct values, and are therefore entitled to be considered independently.

A condition of things not essentially dissimilar can be reported of the next vowel. Its original corresponding short and long sounds would be exactly represented by those heard in the words fill and feel. But the same transition or progression which has waited upon the second vowel has also attended the third. Its proper long sound has now become the name by which we regularly designate the second vowel. The fortunes of i have accordingly been about the same as those of its predecessor e. Here again the genuine short sound has been preserved in its integrity and on a large scale. But the letter is now only occasionally used to denote the long sound it had originally. This employment of it occurs too mainly in comparatively recent words of foreign origin. These have brought with them to a greater or less extent the pronunciation they had in the tongue from which they came. Some of the most common of these words are caprice and police; fatigue and intrigue; profile; machine, magazine, marine, and routine; and antique, critique, oblique, and pique. Once too it belonged to oblige, and even to this day the pronunciation obleege is occasionally heard.

What we call the third vowel is not a vowel, but a diphthong. We can see its sound and real character indicated in the Roman pronunciation of Cæsar, the German kaiser, or in the ae of the Spanish maestro. Against this general movement it can be said that the long and short sounds of the fourth vowel are much nearer their originals. This is by no means true, however, of the fifth. The genuine corresponding long and short sounds of it can be seen represented in the words fool and full. But we now almost universally apply the term “short u” to the neutral sound heard in but and burn. This sound occurs on the most extensive scale. It has, in fact, come to be one of the most common in our pronunciation, as to it all the vowels of the unaccented syllables are disposed to tend. Even the sound of u in accented syllables begins to show occasional traces of this degeneration. Who has not heard that provincial pronunciation of the verb put which gives it the exact value of the initial syllable of putty? With nothing in our orthography to give fixity to orthoepy, there is little limit to the possibilities lying before this so-called “short u” in the way of displacing other sounds.

Let us now summarize the facts of the situation. The primal sound of the first vowel is on the road to complete disappearance. The long sound of the second vowel has usurped the name and in part the proper functions of the first. The long sound of the third vowel has performed a similar office for the second. The third vowel, so-called, is a diphthong. On the other hand, the short sounds of these three vowels—seen in sat, set, sit—continue to exist in their original integrity. All of them are employed on an extensive scale. Furthermore, the regular long and short sounds of u have no longer the prominence they once had in connection with this vowel. To the popular apprehension the idea of it is supplied, as has just been said, by the neutral vowel-sound we call “short u.” This has largely taken the place of other vowel-sounds, and threatens to do so still more in the future.

The confusion in the use of the vowel-signs is itself reinforced by the condition of the alphabet. For the former, indeed, the latter is in no small measure responsible. Behind all the other agencies which have brought about the present wretched condition of our orthography stands out its one most glaring defect. The Roman alphabet we have adopted as our own is unequal to the demand made upon it. The three diphthongs being included in the consideration, we have at a low calculation fifteen vowel-sounds and but five characters to represent them. According to a more common calculation, we have eighteen vowel-sounds to be represented by this limited number. With the consonants we are a good deal better off. The supernumeraries being excluded, there are eighteen single characters for the twenty-four sounds to be denoted.

To make up for this deficiency of letters, two courses lay open to the users of English; rather, two courses were forced upon them. One was to have the same sign represent two or more sounds. This was at best a poor method of relief. Even had it been done correctly and systematically, so far as that result could be accomplished, it could not have failed to be unsatisfactory. It would have been an attempt to impose upon these few signs a burden they were unable to carry. But not even was this imperfect result achieved. Apparently it was not even aimed at. The sounds of the vowels have been so confused with one another that no fixed value can be attached to any vowel-sign. They are often used for each other in the most lawless fashion. So much is this the case that it is frequently impossible to tell from the spelling of a word what is the pronunciation of its vowel, or from the pronunciation of its vowel what is the spelling of the word.

There was another way followed to meet the difficulty. This second method was to make the best of the situation by that combination of vowels, or that combination of vowel and consonant, or of two consonants, to which we have given the name of digraphs. The first of these do not really constitute diphthongs, though such they have sometimes been termed. This method was far more sensible than the preceding. The task of making combinations of letters which should represent only particular sounds would have been, to be sure, a hard one. The lawlessness pervading our vowel-system would doubtless have prevented it from being carried out with thoroughness. But carried out imperfectly, it would have been a distinct improvement upon what we have now. But so far from any attempt having been made to accomplish it on even an imperfect scale, it can hardly be said to have been undertaken at all. There are two instances, indeed, in which such combinations have an invariable or nearly invariable value. One of these is aw, found in such words as bawl and lawn. This digraph never has any other sound than that of the so-called “broad a”—heard, for illustration, in fall and salt. The other is ee, seen in seen itself, as well as in a number of other words. With two or three exceptions, this combination has that sound of the third vowel we now ascribe to the second and call “long e.” But in both these instances the limitation of the digraph to the representation of a single sound was a result of accident rather than of design. These combinations were in truth left to run the same haphazard course which the letters composing them had usually followed. Accordingly, to them extended the lawlessness pervading the vowel-system. As a consequence, the pronunciation of the numerous digraphs became, as we shall see later, as varying and uncertain as that of the single vowels themselves.

We come now to the consideration of specific details upon which have been based the general statements just made. Not by any means all of them. There is no intention here of setting forth an exhaustive enumeration of the facts that could be presented. Even did I possess the phonetic knowledge, which I lack, sufficient to do this properly and fully, the undertaking would have lain outside of my plan. Furthermore, it would hinder the effect of the argument for most persons rather than help it. The mass of detail would be oppressive by its volume, and for that very reason less impressive. Accordingly, I throw out of consideration any representation of the variations of pronunciation to be found in unaccented syllables. In them indistinctness of sound, owing to the inability of our present orthography to denote precise values, has gone beyond that prevailing in the other cultivated tongues of modern Europe. Not only are the vowel-sounds in such syllables pronounced differently by different individuals, they are pronounced differently by the same individual at different times. In particular the precise pronunciation will be apt to vary with the speaker’s rapidity or slowness of utterance. In one case the exact sound will come out with perfect distinctness, in another it will be hard to tell by what vowel it is represented. It is enough to say here of the unaccented syllables that there is a strong tendency, especially in hasty utterance, to give to them generally the sound of that neutral vowel we commonly call “short u.”

It is accordingly in these unaccented syllables that so many were wont to trip in the spelling contests once so popular. It was not unusual to have the very best equipped contestant fail. He attempted to use his reason; to succeed, it was essential to discard that and trust instead to his memory. Take, for illustration, so common a verb as separate. Who, ignorant of the word, could tell from the ordinary pronunciation of it—even when that is reasonably distinct—what is the precise sound heard in the case of the second syllable? Should it be represented by an a or an e? The actual fact has to be learned, not through the agency of the ear, but through that of the eye. This is but a single instance out of hundreds that could be cited where a similar uncertainty must always prevail because the pronunciation cannot act as a clear guide to the present spelling.

In the following pages, therefore, attention shall be directed mainly to setting forth some of the most salient facts which reveal, in a way easily comprehensible, the confusion existing in our present orthography. For this purpose the discussion is intentionally confined almost entirely to those syllables upon which the principal accent falls. In a few instances some syllables will be included upon which rests the secondary accent. In both cases, however, the examples will be selected of words in which the distinction of sound is plainly apparent to all, and easily recognizable. This limits the discussion to but a section of the whole field. But though far from covering the ground, the absolute truth of the general statements about the condition of our orthography will appear distinctly manifest to him who has the patience to wade through the following dreary assemblage of facts, or perhaps it would be more proper to say, the following assemblage of dreary facts. Beginning with the vowel-system, the various letters or combinations of letters will be set forth which are used to indicate the same sound. In a number of instances these signs occur on a very small scale. Accordingly, three examples of every one will be invariably given when the sound heard is represented frequently by the spelling, or at least more or less frequently. When but one or two words are specified, this smaller number will denote that these are all the ordinary ones of that class—exclusive of derivatives and compounds—which are known to exist. At any rate, they are all that are known to exist to the writer. It is not unlikely that examples have been overlooked which will suggest themselves to the reader. We begin with the vowel-system.

III
THE VOWELS

The vowel a demands first attention. The sound of it, heard in father and far, has been spoken of as disappearing. The simple vowel usually represents it, so far as it continues to exist. Other signs, however, are occasionally employed. It is heard in the ua of guard and guardian, in the ea of heart, and also of hearken when so spelled; and finally in England in the e of clerk, sergeant, and a few other words. Once much more common, it has even there steadily given way before the advance of the so-called “short u” sound, occurring in such words as her. In the pronunciation of some it is further represented, for illustration, by the au of haunt and haunch. On the other hand, as contrasted with this declining use, the regular short sound of a, heard in man and mat, is preserved in its fullest vigor. In the large majority of instances it is indicated by the simple letter itself. The exceptions to this representation of it are merely sporadic. Such are the ua of guarantee and the ai of plaid.

But dismissing the consideration of these two sounds of this vowel, take those heard respectively in the words fare, fall, and fate. Let us begin with the first of these. Its sound is denoted in many words by the simple vowel, as can be seen in pare, care, declare. But it is also indicated by ai in pair, hair, stair; by ay in prayer; by e in there and where; and by ei in their and heir. The second of these is the au sound heard in all, warm, want. It is not unfrequently denominated “broad a.” But besides this vowel the sound is further represented by o in such words as oft, loss, song; by au in daub, haul, taught, and the like; similarly by aw in saw, drawn, bawl, and numerous others; by oa in broad; and by ou in sought, thought, bought.

It has already been pointed out that the so-called long sound of a does not strictly belong to it; that it is really an e sound. But as it has imposed its name upon the vowel, it is properly to be considered with it in any treatise which appeals to the general public. Its most usual representative is the letter itself, seen in pale, pane, page, and in scores of words in which the presence of an unpronounced final e has come to indicate generally, though not invariably, that the preceding vowel is long. But then again it is represented by ai in pail, pain, exclaim; by ay in lay, pay, day; by ea in great, steak, break; by ei in veil, vein, heinous; and by ey in they, obey and survey. In the interjection eh the vowel has for once its original sound. Again there are two instances in which a digraph with this sound occurs in but a single case. These two are the ao of gaol and the au of gauge.

In the case of the first of these words there were two ways of spelling it which existed from the fourteenth century. These are gaol and jail. The first form comes from the dialect of Normandy, the second from that of Paris. Both have been in use from the beginning. About both there has been to some extent controversy, at least in the past. The New Historical Dictionary, which contains a full history of the origin and use of these two forms, gives us a quotation bearing upon this point from Roger L’Estrange’s translation of the Visions of Quevedo. In this version, which appeared in 1668, English allusions were not unfrequently introduced. In one instance men are represented as being in a state of rage because they cannot come to a resolution as to whether they ought to say Goal (sic) or Jayl. Gaol is still the official form of the word in England. That fact has mainly contributed to its maintenance in literature, so far as it continues to be used. In the United States jail is both the official and the literary form. But the spelling gaol has to some a peculiar attraction of its own. Not a single letter in it save the final l is of use in indicating with certainty its right pronunciation. In truth, the orthography almost enforces a wrong one. There are those to whom this fact is the highest recommendation it can have.

The second word has varied between the spellings gauge and gage almost from its very entrance into the language in the fifteenth century. One gets the impression that there was a time when the latter was the preferred form. But with our present knowledge no statement of this sort can be made positively. “You shall not gage me by what we do to-night,” says Gratiano to Bassanio in The Merchant of Venice. Modern editions, in defiance of the original, print gauge; for the folio and both the early quartos agree in having gage. Shakespeare’s use seems to be nothing but another illustration of his perverse preference for the so-called American spelling, or the American preference for Shakespeare’s spelling, just as one chooses to put it. Such an anomalous form as gauge proved at times too much for the tolerance of the orthographically much-enduring Englishman. Even him it has struck as peculiarly objectionable. So in the eighteenth century he set out to remove this particular blot upon the speech. But as he was in nowise tainted with the virus of reform, he exhibited the usual incurable aversion to having the spelling bear any relation to the pronunciation. Accordingly, he refused to take the natural as well as time-honored course of dropping the unnecessary and misleading u. Instead, he reversed the order of the letters of the digraph. The au became ua.

There have been in modern times men who advocated this method of spelling the word with all that fervor of faith which is so frequent an accompaniment of limited knowledge. On this point, for instance, the late Grant Allen felt called upon to bear his testimony. He was wont to make his novels a vehicle for conveying his linguistic views as well as those pertaining to religion, society, and politics. “Cynicus replied, with an ugly smile,” he wrote, “that nobody could ever guage anybody else’s nature.”[16] Then, with what might fairly be called an ugly smile of his own, Allen added in a parenthesis, “not gauge, a vile dictionary blunder.” There was no apparent reason for this lexical outburst; there was certainly no proof vouchsafed of the justice of the assertion. As the originals of the word were the Old French noun gauge and the verb gauger, it is hard to see how dictionaries could be held responsible for blunders, if blunders they were, which foreigners had perpetrated centuries before. There is, in truth, as little etymological justification for guage as there is phonetic for gauge. Gage, if it were not the most common way of spelling the word during the Elizabethan period, was certainly a common one. It is now, on the whole, the preferred form in the United States. Except in the nautical term weather-gage, the u is very generally retained in England. This is doubtless due to the desire of gratifying the ardent enthusiasm pervading the toiling millions of Great Britain for spellings which remind them of the Old French originals, from which were derived the words they employ.

In the case of the second vowel, the short e sound is properly shown in a large number of words of which let, felt, bed may be taken as representatives. These are all phonetically spelled. No educated man who saw them for the first time would have any hesitation about their pronunciation. Such a condition of things tends to chasten the feelings of that class of persons, not inconsiderable in number, who think it distinctively to the credit of the spelling that it should get as far away from the pronunciation as possible. They may be consoled, however, by the fact that this same sound is represented by a in any and many; by ea in a large number of words, such as health, endeavor, weather; by ai in said and again; by ay in says; by ei in heifer and nonpareil, and by eo in jeopard and leopard. There are those who give this short sound to leisure, rhyming it with pleasure, as did Milton,[17] instead of the more common long sound heard with us. Indeed, it is noticeable that preference is given to the former in the New Historical English Dictionary, though that pronunciation is absolutely ignored in some of the best American ones. The compilers of these last may have been touched by Walker’s pathetic plea for the long sound. “Leisure,” he wrote, “is sometimes pronounced as rhyming with pleasure; but in my opinion very improperly; for if it be allowed that custom is equally divided, we ought, in this case, to pronounce the diphthong long, as more expressive of the idea annexed to it.”

Any and many are now the only two words where a has the sound of short e. At one time it was heard in others, and was not unfrequently so represented in literature. It lingers, too, in some instances, and even, indeed, flourishes in spite of all the efforts of education to extirpate it. The present authorized pronunciation of catch, instead of ketch, is one of the comparatively few triumphs gained by the written word over the spoken. In days when devotion did not exist to orthography irrespective of the purpose it was designed to fulfil, the a assumed the spelling of e along with its sound. The earlier cag, for illustration, has been abandoned for the pure phonetic spelling keg. Apparently no serious harm has befallen the language in consequence. Even more distant from the remote Latin original, canalis, denoting the home of canis, ‘the dog,’ is the form kennel. This turns its back upon its primitive, and contents itself with simply representing the pronunciation. So much are we the creatures of habit and association in the matter of spelling that the most ardent believer in the doctrine of basing orthography upon derivation could in neither of the cases just mentioned be persuaded to revert to the form nearest to that in the original tongue.

The sound to which we give the name of “long e” belongs strictly, as has been pointed out, to i. A few of the words have also been given in which it still continues to be so indicated.[18] There are certain conditions under which it is represented by the simple letter itself. One is when it alone constitutes an accented syllable, as in equal, era, ecliptic. Another when it ends a monosyllable or an accented syllable, as in he, be, regal, cohesion. It appears finally with a good deal of frequency in words in which the sound of the simple vowel is lengthened by the artificial device of an appended mute e, as in theme, precede, complete. This last word, it may be said in passing, was once often spelled compleat. But as the letter itself represents much more usually the short sound of the vowel, the long sound has come to be indicated often by various digraphs. Of these, two are particularly prominent. One of them is ee, seen in a large number of words, such as meet, thee, proceed. The second digraph is ea, found in bean, meat, eagle, and a host of others. But the sound is not limited to these two combinations. It is represented by ei in receive, conceit, seize; by ie in believe, chief, fiend; by ey in key; by eo in people; by ay in quay; by æ in ægis, pæan, minutiæ; and several other words not fully naturalized.

Once, indeed, this last method of indicating the sound was far more common. In many instances it has been supplanted by the simple e. It was not till a comparatively late period that such spellings as era and ether drove out in great measure the once prevalent æra and æther. As æ has with us strictly but one sound, the change cannot, from all points of view, be deemed an improvement. In the case of an unknown word first brought to the attention, no one could now be positive, under certain conditions, whether the vowel should be treated as long or short. Take, for illustration, encyclopedia, once often spelled encyclopædia. He who sees the word for the first time is as likely to pronounce the antepenultimate syllable pĕd as pēd. He certainly could not tell from the orthography employed how this particular syllable should be sounded. Still, for much more than a century the tendency of the users of the language has been steadily directed toward the discarding of the æ in all cases. As long ago as 1755 Dr. Johnson, in his dictionary, recommended its disuse. “Æ,” he wrote, “is sometimes found in Latin words not completely naturalized or assimilated, but is by no means an English diphthong, and is more properly expressed by single e, as Cesar, Eneas.” Dr. Johnson was hostile to spelling reform; but he could venture to sanction a spelling of these two Latin proper names, at which even the average spelling reformer would shudder.

Fortunately for those of us who believe that spelling exists for the sake of indicating pronunciation, the sound of short i, one of the most common vowel sounds in the language, is almost always represented by the letter itself. The exceptions are few, comparatively speaking. The only sign to take its place in any body of words sufficiently numerous to be entitled a class is y, as seen in syntax, abyss, system, and other words, generally of Greek origin. The instances where different signs are employed are purely sporadic. Most of them, however, are for various reasons remarkable. The sound is represented by e in the name of the language we speak and of the country where it came into being. It is further represented by the e of pretty, by the o of women, by the u of busy and business, by the ie of sieve, and by the ui of guild and guilt and build. Once in the speech of most men, and now in that of many, it is given to the ee of been, and regularly to that combination as found in breeches.

Gild is a variant spelling of guild, and represents the earlier form. The ui of the two additional examples given ought to be a saddening spectacle to the devout believer in derivation as the basis of orthography. The original form of guilt was gylt. So it remained with various spellings—of which gilt was naturally the most common—until the sixteenth century. But there was also an allied form, gult. These two undoubtedly represented distinct and easily recognizable pronunciations of the word. They were at last combined so as to create a spelling, of the pronunciation of which no one could now be certain until he was told. This did not take place on any scale worth mentioning until the latter part of the sixteenth century, though the combination had occasionally been seen much earlier. Essentially the same thing can be said of build. It originally appeared in various ways, of which byld, bild, and buld were the prominent types. At the end of the fifteenth century the practice began of recognizing both forms by writing build or buyld. In a measure this doubtless represented a then existing shade of pronunciation. The spelling, once established, has continued since. No one ever thinks of pronouncing the u; perhaps no one has ever thought of it since the combination was formed. Yet there is no question that intense sorrow would be occasioned to a certain class of persons were they to be deprived of the pleasure of inserting in the word this useless and now orthoepically misleading letter.

The so-called “long i” ought strictly to be treated under the diphthongs; but as it is popularly associated in the minds of men with the simple vowel, its diphthongal sound will be considered at this point. Its most usual representative is the letter itself. This presents little difficulty in the pronunciation if the words end with a mute e, as in mine, desire, bite. The distinction between thin and thine, for instance, is then easily made. But when it comes to such words as mind, child, and pint on the one hand, and lift, gild, and tint on the other, there is nothing in the spelling to indicate with certainty how the i of these words should be sounded. As no general rule can be laid down, the pronunciation of each has in consequence to be learned by itself. This uncertainty was perhaps one of the causes which led to the transition of the diphthongal sound of i in wind to the short sound wĭnd, which so aroused the wrath of Dean Swift. But besides i the sound is also indicated by the y of type, ally, thyme, and a number of words derived from the Greek; by ie, especially in monosyllables, such as die, lie, and tie; by ye in the noun lye; by ei in height and sleight, and according to one method of pronunciation in either and neither. It is further represented by the ai of aisle, by the ey of eye, and by the uy of buy.

The third vowel now demands attention. Orthoepists contend that there is no genuine short o in English utterance. Without entering into a discussion of this point, it is sufficient to say that the two sounds of the letter, which are ordinarily designated as short and long, are represented respectively in the words not and note. The former sound remains fairly faithful to this vowel. It is hardly indicated by any other sign. The a of what, squad, quarry is about the only one to take its place. Very different is it with the long sound heard in note. This is far from confining itself to any single letter. In no small number of words it is represented by oa, as in boat, groan, coal; by oe, as in foe, toe, hoe; by ou, as in pour, mould, shoulder; or again by ow, as seen in crow, snow, show. Less common, but still to be met with, is this sound heard in the combination ew, as seen in sew; as well as in shew and strew, as these words were once regularly and are now occasionally spelled; in oo, as in door and floor; in eau, in beau, bureau, and flambeau; and in the eo of yeoman.

This last word was once spelled at times yoman and at times yeman. These forms doubtless represented the two ways of pronouncing it that existed. The Toxophilus of Roger Ascham, for illustration, was dedicated to the use “of the gentlemen and yomen of Englande.” But the sound of the vowel of the first syllable wavered for a long period between the long o and the short e. Ben Jonson, in the earlier part of the seventeenth century, observed of the word that “it were truer written yĕman.” In the latter half of the eighteenth century Doctor Johnson tells us that the eo of this word “is sounded like e short.” This was the view taken by perhaps the larger number of orthoepists, who immediately followed him. In spite of them the o pronunciation has triumphed. It has shown, however, a tender consideration for its defeated rival by allowing it to lead a useless existence in the syllable in which, in the utterance of many, it once represented the actual sound.

The corresponding short and long sounds of u are seen in the words full and rule. But o, either singly, or in combination with other letters, is a favorite way of indicating both. The short sound of this vowel, which is far from common, is represented by the o of bosom, woman, wolf; by the oo of good, foot, stood; by the ou of could, would, should. On the other hand, the corresponding long sound is also represented by the o of move, prove, lose; by the oe of shoe and canoe; by the oo of too, root, fool; by ou in such words as uncouth, routine, youth, and a number of others derived generally from the French. There has been and still is something of a tendency on the part of the users of language to change the long sound of oo into its short one. Oliver Wendell Holmes, in his poem of Urania, represents Learning as giving a lesson on propriety of pronunciation. Among other points considered, occurred the following observations:

She pardoned one, our classic city’s boast,
That said at Cambridge mŏst instead of mōst,
But knit her brows and stamped her angry foot
To hear a Teacher call a rōot a rŏot.

This is perhaps as good an example as can be furnished of the waste of time and labor imposed by our present orthography in mastering distinctions of sounds in words when there is nothing in the sign employed to indicate which one is proper. To men not given up to slavish admiration of our present spelling, it would seem that Learning, instead of stamping her foot, would have been much more sensibly engaged in using her head to devise some method by which one and the same combination of letters should not be called upon to represent two distinct sounds in words so closely allied in form as foot and root; or distinct sounds in words with the same ending as toe and the shoe that covers it.

Another way of indicating the long sound of this vowel is either by the simple letter itself or by it in combination with other letters. For instance, it is represented by ue in such words as true, avenue, pursue; by ui in fruit, bruise, pursuit; by eu in neuter, deuce, pentateuch; by ew in brew, sewer, lewd; and by ieu in adieu, lieu, purlieu. But there is a peculiarity in the words containing this vowel, the consideration of which involves too much space to have little more than a reference here. We all recognize the difference of the sound of u as heard respectively in fortune and fortuitous, in annual and annuity, in volume and voluminous, in penury and penurious. In the first one of each of these pairs of words a y-element is introduced into the pronunciation; in the second the u has its absolutely pure long sound. Nor is this introduction of the y-element limited to the letter when used alone. We can find it exemplified in the ue of statue, value, tissue; in the eu of eulogy, euphony, Europe; in the ew of ewe, hew, few. This iotization, as it is called, is especially prevalent in words with the termination ture, as nature, furniture, sculpture, and agriculture. Now and then some one is heard giving, or attempting to give, to the u of this ending the pure sound; but such persons are usually regarded as possessed of “cultoor” and not culture.

The only word of this special class in which such a method of pronunciation can be said to have attained any recognition whatever is literature. The word itself is an old one in our speech. Once, however, it meant merely knowledge of literature. It did not mean that body of writings which constitute the production of a country or of a period. This sense of it, now the most common, is comparatively modern. The earliest instance I have chanced to meet of it—though it was doubtless used a good deal earlier—is in the correspondence of Southey and William Taylor of Norwich. There it occurs in a letter belonging to the year 1803, in which Southey tells his friend that he was expecting to undertake the editorship of a work dealing biographically and critically with “the history of English literature.”[19] Still, the pronunciation just mentioned of this word, differing as it does from the others of the same class, must even then have been occasionally heard. It was certainly made the subject of comment by Byron. He somewhere speaks—I have mislaid the reference—of a publisher who was in the habit of talking about literatoor. This peculiar pronunciation still comes at times from the lips of educated men.

But the regular long and short sounds of u yield in frequency of occurrence to that sound of it heard in but and burn. In common speech this has usurped with us the title of “short u.” By orthoepists it itself is divided into a long and a short sound, according as it is or is not followed by an r. Into it, as has been pointed out, the pronunciation of all unaccented syllables tends to run. Hence, in the case of these, there has come to exist the greatest possible variety of signs by which it is indicated. But even in the accented syllables there is a sufficient number of different ones to arrest the attention. Naturally the most usual representative of it is the vowel from which it has taken its name. But it is far from being limited to this sign. Its short sound is further represented by the o of such words as love, dove, and son; similarly by the ou of double, touch, and young; and by the oo of blood and flood. In vulgar speech soot would have to be added to the last two. Furthermore, it is represented by the sporadic example of the oe of does. The long sound runs through a still wider range of examples. Words containing it but denoted by various signs could be given by the score. It is represented by all the vowels except the first. The e of her, were, fern, stands for it. So does the i of fir, bird, virgin. So does the o of work, worship, worth. It is likewise largely represented by ea in such words as heard, learn, search; by ou in scourge, journal, flourish, and no small number of others containing this particular sign. In the single instance of tierce the sound is denoted also by ie. Were its use in unaccented syllables indicated, this list of signs would be largely extended. As it is, it will be seen that nine is the number employed in accented syllables to represent it.

So much for the simple vowels. We come now to the three diphthongs. The first of these, which is made up of the sound of the a of father and that of the e of they, has already been considered in treating what is called “long i.” Eight signs were given by which it was denoted.[20] This wealth of representation does not belong to the two other diphthongs. There are but two signs by which the sound of the second is indicated. These are the ou of south, found, about, and the ow of now, town, vowel. The third diphthong again has but two signs, the oi of boil, point, spoil, and the oy of boy, joy, destroy. Many of the words in which oi appears had once the pronunciation of the first mentioned diphthong. To the truth of this both the rymes of the poets and the assertions of the early orthoepists bear ample testimony. The statement is still further confirmed by the fact that the sound still lingers, or, rather, is prevalent, in the speech of the uneducated, the great conservators of past usage. The words given above as illustrative of this sign of the diphthong would have been pronounced by our fathers bīle, pīnt, spīle. So they are still pronounced by the illiterate. In one word, indeed, this sound has not passed entirely from the colloquial speech of the cultivated either in England or America. Roil is not merely heard as rīle, but is not unfrequently found so printed.

IV
THE DIGRAPHS

Up to this point we have been engaged in making manifest the numerous different ways in which the same vowel-sound is represented in our present orthography. Necessarily a reversal of the process would present an equally impressive showing, for examples just as impressive would make manifest how the same sign adds to the further confusion of English spelling by denoting a number of different vowel-sounds. But there is a limit to the endurance of the reader, to say nothing of that of the writer. Furthermore, there is little need of this addition in the case of the vowels. The facts about to be furnished will be more than sufficient to satisfy any demand for illustrations of the extent to which the same sign has been made to indicate a wide variety of different sounds, though in the sporadic instances the examples already given must be repeated. For we come now to the consideration of those combinations of letters, numerous in English spelling, to which has been given the name of digraphs. They are sometimes made up of a union of vowels, sometimes of a union of a vowel and a consonant, sometimes of a union of two consonants.

I have already adverted to the fact that had there been any system established in the employment of these combinations of letters, and had each of them been made to represent unvaryingly one particular sound, some of the worst evils of English orthography would have been largely mitigated, and in certain cases entirely relieved. But this was not to be. The opportunity of bringing about regularity of usage in the employment of these signs was either not seen, or if seen was not improved. The same variableness, the same irregularity, the same lawlessness which existed in the representation of the sounds of the vowels and diphthongs came to exist in the case of the digraphs also. They consequently did little more than add to the confusion prevailing in English orthography, and became as valueless for indicating pronunciation as are the single letters of which they are composed.

To this sweeping statement there are two partial exceptions. The first is aw. This is one of several representatives of the so-called broad sound of a heard in ball and fall. Whenever that digraph appears, its pronunciation is invariably the same. No such absolute assertion can be made of the digraph which represents the sound of “long e.” This is the combination ee. There are but two exceptions in common use to the pronunciation of it just given. The first is the word breeches. Its singular has the regular sound. The pronunciation as short i in the plural—used, too, there in a special sense—may perhaps be due to an extension to this form of that tendency, so prevalent in English speech, on the part of the derivative, to shorten the vowel of the primitive. The other is the participle been of the substantive verb. In usage the pronunciation of this word has long wavered and still wavers between the sounds heard respectively in sin and seen. Of this variation there will be occasion to speak later in detail.

These exceptions, however, affect but a limited number of words. They are hardly worth considering when their regularity is put in contrast with the irregularities of the other combinations. Let us begin with the digraph ai. Ordinarily it has the sound we are accustomed to call “long a,” as can be seen in fail, rain, and paid. In pair, fair, hair it has another sound. In said, again, against it has the sound of short e. In aisle it has the diphthongal sound called by us “long i.” This word furnishes an interesting illustration of the way in which much of our highly prized orthography came to have a being. Its present spelling is comparatively recent. Doctor Johnson recognized in it the lack of conformity to any possible derivation. He adopted it on the authority of Addison, though with manifest misgiving. He thought it ought to be written aile, but in deference to this author he inserted it in his dictionary as aisle.

“Thus,” he said, “the word is written by Addison, but perhaps improperly.”[21] Johnson’s action was followed without thought and without hesitation by his successors. There is no question, indeed, as to the impropriety of the present spelling from the point of view of both derivation and pronunciation. Equally there is no doubt as to the impropriety of its meaning from the former point of view. It came remotely from the Latin ala, ‘a wing.’ Therefore, it means really the wing part of the church on each side of the nave. In this sense it is still employed. But since the first half of the eighteenth century it has been made to denote also the passage between rows of seats. Strictly speaking, this is a particularly gross corruption, though, like so many in our speech, it has now been sanctioned by good usage. The proper word to indicate such a sense was alley, corresponding to the French allée, ‘a passage.’ This was once common and is still used in the North of England. Aisle itself was formerly spelled ile or yle. Confusing it with isle, originally spelled ile, men inserted an s about the end of the seventeenth century. Later an a was prefixed under the influence of the French aile. It was thus that this linguistic monster, defying any correct orthography or orthoepy, was created. In any sense of it the s is an unjustifiable intrusion, representing as the word does in one signification the Latin ala, ‘a wing,’ and in the other the French allée, ‘a passage.’

There is another word containing this digraph which illustrates vividly the uncertainty of sound caused by the present spelling. This is plait, both as verb and substantive. About its pronunciation usage has long been conflicting. “Plait, a fold of cloth, is regular, and ought to be pronounced like plate, a dish,” said Walker, at the beginning of the eighteenth century. “Pronouncing it so as to rhyme with meat,” he added, “is a vulgarism, and ought to be avoided.” So say several later English dictionaries. So say the leading American ones. Webster, indeed, concedes that the pronunciation denounced by Walker is colloquially possible. It therefore does not necessarily relegate the user of it to the ranks of the vulgar. Now comes the New Historical English Dictionary, and gives the word not merely three distinct pronunciations, but holds up as only really proper that which has failed to gain the favor of most lexicographers. It is the one found “in living English use,” it says, when the word has the sense of ‘fold.’ Further we are assured that with this signification it is ordinarily written pleat. This would tend to justify still more the ryme with meat, which so shocked Walker. Then in its second sense of a ‘braid of hair or straw’ we are told that it has the sound of a in mat. This leaves the pronunciation of plait to ryme with plate hardly any support to stand on. It has merely the distinction of being mentioned first; but it is denied a real existence as a spoken word. Nothing could better illustrate the unlimited possibilities opened by our present orthography for discussions of propriety of pronunciation about which certainty can never be assured. All statements about general usage, no matter from what source coming, must necessarily be received with a good many grains of allowance, if not with a fair proportion of grains of distrust—at least, whenever our orthoepic doctors disagree. Do the best the most conscientious investigator can, he can never make himself familiar with the practice of but a limited number of educated men who have a right to be consulted. His conclusions, therefore, must always rest upon a more or less imperfect collection of facts.

The digraph ay is naturally subject to the same influences as ai. It is, however, much less used save at the end of words. Grief has, indeed, been felt and expressed, even by devout worshippers of our present orthography, at the arbitrary change of signs made in the inflection of certain verbs, like lay, pay, say. These, without any apparent reason for so doing, pass from the digraph ay in the present to ai in the preterite. Naturally there is no objectionable uniformity in the practice. That might tend to render slightly easier the acquisition of our spelling. Accordingly, lay and pay and say have in the past tense laid, paid, and said, while verbs with the same termination, such as play, pray, delay, have in this same past tense the forms played and prayed and delayed. Stay uses impartially staid and stayed. Much dissatisfaction has been expressed at the “wanton departure from analogy,” as it has been called, which has been manifested by the words of the first list given. As the characteristic of our spelling everywhere is a wanton departure from analogy, it hardly seems worth while to find fault with this particular exhibition of it.

In quay the digraph has the entirely distinct sound of “long e.” Of this word it may be added that the spelling is modern while the pronunciation is ancient. Originally it appeared as key or kay—of course, with the usual orthographic variations. In the earlier half of the eighteenth century, under the influence of the French quai, the present form of the word came in; toward the end of the century it had become the prevailing form. This gave the lexicographer, Walker, an opportunity to display his hostility to any sort of spelling which should engage in the reprehensible task of aiming to indicate pronunciation. Such a proceeding was in his eyes a radically vicious course of action. In the entire ignorance of the original form of the word he remarked that it “is now sometimes seen written key; for if we cannot bring the pronunciation to the spelling, it is looked upon as some improvement to bring the spelling to the pronunciation—a most pernicious practice in language.”

Key, as the spelling suggests, had originally the sound of ey in they and obey; later it passed into the sound of “long e.” This it has transmitted to its supplanter. In the scarcity of rymes in our tongue, it is always a little venturesome to infer from the evidence of verse the past pronunciation of words which have with us the same termination, but different sounds. This imparts a little uncertainty to the treatment of quay in two passages containing the word which are given by the New Historical Dictionary.

But now arrives the dismal day
She must return to Ormond-quay,

says Swift in his poem of Stella at Wood-Park. Does the ryme here represent an attempt to conform the pronunciation to the spelling? More likely it represents the survival of a pronunciation once more or less prevalent. The second extract from In Memoriam is under the circumstances more striking:

If one should bring me the report
That thou hadst touched the land to-day,
And I went down unto the quay,
And found thee lying in the port.

It certainly looks as if in this passage Tennyson had set out to make the pronunciation conform to the spelling.

Our next digraph is ea. This has a choice variety of sounds to represent. Most commonly it receives the pronunciation of “long e.” Of the scores of examples containing it, beast, hear, and deal may be taken as specimens. But while this is its most frequent sound, it is far from being the only one. Its most important rival is that of short e, which can be found in no small number of words like breath, breast, weather. In these and all other like cases the second vowel is absolutely superfluous as regards pronunciation. The unnecessary letter is in some instances due to derivation; in others it exists in defiance of it—as, for instance, in feather and endeavor. Its insertion was doubtless due to an attempt to represent a sound which is no longer heard in these words. In a large number of instances they were once spelled without the now unpronounced letter.

Common also is a third sound of this digraph—the one we call “short u.” It is heard in heard itself, in earth, in early, in learn, in search, and in a number of other words in which ea is followed by r. There is a fourth sound of it which may be represented by bear, swear, tear. A fifth sound of it occurs in the words heart, hearth, and hearken. Again, a sixth sound of it is represented by such words as great, break, steak. In all these cases it will be observed that certain of these words have in the course of their history tended to pass from one pronunciation of the digraph into another. Sometimes they have for a long time wavered between the two. Hearth, which contains the fifth sound just assigned to the combination, was often made to ryme with words containing the third sound, represented by earth. According to the New Historical Dictionary, this is true now of Scotland, and of the Northern English dialect. It is true also of certain parts of the United States, or, at any rate, of certain persons. It seems also to have been the pronunciation of Milton.

Far from all resort of mirth
Save the cricket on the hearth

are lines found in Il Penseroso. So also great once had often the first sound here given to the digraph, as if it were spelled greet. Both this sound as well as the one it now receives were so equally authorized in the eighteenth century that Dr. Johnson triumphantly cited the fact as a convincing proof of the impossibility of making a satisfactory pronouncing dictionary, just as we are now told that we cannot have a phonetic orthography because men pronounce the same word in different ways.

The digraph ee having already been considered, we pass on to ei. Its most frequent sound is that heard in such words as rein, veil, and neighbor. But it has also the sound of “long e” in conceit, seize, ceiling, and a few others. In heir and heiress and their it has the sound of a in fare. In height and sleight it has the sound of “long i.” In heifer and nonpareil it has the sound of short e. The allied digraph ey has no such range of sounds. In accented syllables it represents only the first one given to ei, as can be seen in they, grey, and survey. Key, with the sound of “long e,” seems to be the solitary exception.

It is already plainly apparent that there is nothing in the character of our present spelling to fit it to serve as a guide to pronunciation, the very office for which spelling was created. But its worthlessness in this respect, with the consequent uncertainty and anxiety attending the use of it, forms in the case of two words containing the digraph ei, one of the most amusing episodes in the history of English orthoepy. In modern times their pronunciation has given rise to controversy and heart-burnings as bitter as the matter itself is unimportant. These words are either and neither. Were they to adopt the most common pronunciation of the digraph they would have the sound heard in such words as eight, vein, and feint. This, in truth, they once had. To indicate that fact they have occasionally been written ayther and nayther. But this pronunciation, outside of Ireland at least, had largely disappeared by the latter part of the eighteenth century. So far as many orthoepists were concerned, it was ignored entirely. Those who mentioned it often accorded it scant favor. The affections of lexicographers were long divided between the sounds heard in receive and deceit, and that heard in height and sleight. For the former there was a very marked preference. Most of them did not even admit the existence of the “long i” sound; those who did, gave it generally a grudging recognition. The various pronunciations prevailing in the latter part of the eighteenth century were specified by Nares in his Elements of Orthoepy. “Either and neither,” he wrote, “are spoken by some with the sound of long i. I have heard even that of long a given to them; but as the regular way is also in use, I think it is preferable. These differences seem to have arisen from ignorance of the regular sound of ei.” As the regular sound of ei, if any one of them is entitled to that designation, is heard in such words as skein and freight, one gets the impression that Nares himself was ignorant of what it was.

Walker, the orthoepic lawgiver of our fathers, distinctly preferred the “long e” sound of either and neither. Both the practice of Garrick and analogy led him to maintain that they should be pronounced as if ryming “with breather, one who breathes.” He was compelled, however, to admit that the “long i” sound was heard so frequently that it was hardly possible in insist exclusively upon the other. He did the best he could, nevertheless, to ignore it and thereby banish it. While in the introduction to his dictionary he recognized the existence of both sounds, in the body of his work that of “long e” was the only one given. In this course he was followed by his reviser, Smart, who succeeded to his name, and up to a certain degree to his authority. Smart went even further than his predecessor. He was apparently ignorant of the fact—he certainly ignored it—that any other pronunciation of these words than that of “long e” was known to the English people. But in spite of its defiance of analogy and of the hostility of lexicographers, the sound of “long i” continued to make its way. The fact has sometimes excited the indignation of orthoepists. Yet it is hard to understand how any one who cherishes the vagaries of English spelling should get into a state of excitement about the vagaries of its pronunciation.

Neither the digraph eo nor eu is found often. The first, however, improves fully the opportunity presented of making it difficult, if not impossible, for the learner to get any idea of the pronunciation from the spelling. In people it has the sound of “long e”; in leopard and jeopard it has the sound of short e. In yeoman again it has the sound of long o. Eu has practically the same sound as ew, as can be exemplified in feud and few. This last digraph, however, represents the long sound of u as well as that in which iotization precedes the vowel. The difference in the pronunciation of drew and dew will make manifest the contrast. There is always a tendency, however, for the digraph to pass from the latter sound to the former in a tongue in which there is nothing in the orthography to fix a precise value upon the sign indicating both. “According to my v’oo,” Oliver Wendell Holmes, in his Elsie Venner, has one of his characters saying. “The unspeakable pronunciation of this word,” he adds in a parenthesis, “is the touchstone of New England Brahminism.” In another place in the same novel he still further enforces this point. “The Doctor,” he wrote, among his other recommendations to the hero, says to him, “you can pronounce the word view.” And yet in it the iotization is plainly indicated by the vowel itself, while in such words as hew and few and new there is nothing to fix definitely the sound. Finally, it remains to say of this digraph that shew and strew, two verbs once spelled with it, have now become show and strow, a form more in accordance with their pronunciation. There is no particular reason why sew should not follow their example in substituting an o for an e.

The digraph ie is represented but by three vowel sounds. The most common one is that of “long e”—seen, for example, in chief, grieve, believe. But the sound of “long i” is heard in no small number of words, especially monosyllabic words ending in ie, such as lie, die, tie. In one instance it has the sound of short e. Accordingly, in it the first vowel is distinctly superfluous. This is the word friend. Its Anglo-Saxon original is freónd, just as that of fiend is feónd or fiónd. One of the small jokes of the opponents of spelling reform is a professed unwillingness “to knock the eye out of a friend.” Disparaging remarks have been made about this as an argument—as it seems to me, with no justification. Compared with most of the objections brought against the efforts to wash the dirty face of our orthography and make it decently presentable, this particular argument against dropping the i out of friend is, as I look at it, the strongest that has been or can be adduced. It reminds one, indeed, of the objection the French writer made to the dropping of the h out of rhinoceros. The animal would lose his horn and become nothing more than a sheep.

As a matter of linguistic history, however, it was not until late in the sixteenth century that the i, though found long before, appeared in the word friend to an extent worth considering. There were several ways in which it had been spelled previously. Of these frend was naturally a common one in days when the belief still lingered that the office of orthography was to represent pronunciation and not to get as far away from it as possible. Take, as an illustration, the treatise entitled The Schoolmaster of the great English scholar, Roger Ascham. This appeared in 1570. In it the word friend occurs just twenty-five times. It is regularly spelled frend, with the exception of one instance, where the intruding i is found. So also frendly is invariably the form of the adjective, and frendship that of the derivative noun.[22]

Oa, the next digraph in order, comes very near attaining the distinction of being represented by a single sound. It occurs in a fairly large number of words which can be represented by oar, coat, loaf. It is saved, however, from the reproach of regularity by having the sound of the a of “ball” in the words broad, abroad, and groat. Oe is not so common, but, like its reverse eo, what it lacks in number of words it makes up in variety of pronunciation. In foe, hoe, and toe it has the sound of long o. In canoe and shoe it has the sound of long u. In these instances it forms the termination of words. Not so in does, where it has the sound we call “short u.” The use of this digraph, like that of ae, has been much restricted. For instance, the word we now spell fetid was once generally spelled fœtid. So, in truth, it continued to be till the nineteenth century. The digraph, indeed, still lingers in the name of the drug asafœtida, though in the instance of this word the long sound has given way to the short. Not unlike, in some particulars, has been the fortune of certain other terms. Take, for instance, the word economy. Its remote Greek original began with oi, which in English, as in Latin, appeared with the form œ, and sometimes erroneously æ. For these was found occasionally the simple e. In the nineteenth century this last displaced the two others, and gave to the first syllable the present standard form. One of the results, however, of this sort of substitution is that no one seems to be certain whether he ought to pronounce the initial e of economic as long or short.

The ordinary sound of oo, the next digraph to be considered, is that of long u, as we see it in moon, soon, food. But there are about half a dozen words—throwing derivatives out of consideration—in which it has the short sound of u. The difference can be plainly observed by contrasting the pronunciation of the digraph in the two words mood and wood. Furthermore, oo is to be credited with two more sounds. One is that of the “short u” seen in blood and flood. The other is the long sound of o in door and floor, anciently spelled dore and flore. Dore, for instance, can be found in Chaucer, Shakespeare, Milton, and even as late as Bunyan.

The digraph ou is perhaps the banner sign for the frequency of its occurrence and the variety of sounds it indicates. As it appears most commonly, it is a genuine diphthong, as seen in such words as loud, sour, mouth. But there is another large body of words in which the sign has a sound essentially distinct. It can be observed in such words as group, youth, tour. It gives one a peculiar idea of the worth of English orthography as a guide to pronunciation that in thou, the singular of the pronoun of the second person, ou has one value, and in its plural, you, it has a value altogether different. The same observation is true of the possessives our and your. There are two or three words in which these two signs have had for a long period a struggle for the ascendancy. Take the case of the substantive wound. One gets the impression from poetry that in this word the ou constitutes a genuine diphthong. There is no question that it rymes regularly with words containing the diphthongal sound here given. Perhaps that was a necessity; it had to ryme with such or not ryme at all. Still, the verse seems pretty surely to have represented the common pronunciation. In the couplets of Pope, the poetic authority of the eighteenth century, it is joined, for instance, with bound, found, ground. Yet this same pronunciation was unequivocally condemned by Walker at the end of the same century. “To wound,” he writes, “is sometimes pronounced so as to rhyme with found; but this is directly contrary to the best usage.”

This same uncertainty in the pronunciation of words in consequence of the uncertainty of the pronunciation of the signs employed to represent it may be further exemplified in the case of the noun route. Unlike wound, which is a pure native word, this is of French extraction. Following the analogy of most of the words so derived, it ought to have the second sound given here to the digraph. Yet it not unfrequently receives that of the first. Thus Walker graciously tells us that it is often pronounced so as to ryme with doubt “by respectable speakers.” A far more interesting case is that of pour. The majority of eighteenth century orthoepists—Johnston, Kenrick, Perry, Smith, and Walker—pronounced the word so as to ryme with power. Spenser so employed it. So did Pope, more than a century later. In the only two instances he uses the word in his regular poetry at the end of a line it has this sound. In his Messiah occurs the following couplet:

Ye Heavens! from high the dewy nectar pour:
And in soft silence shed the kindly shower.

Walker, indeed, declared unreservedly that the best pronunciation of it is “that similar to power.” Nares alone among eighteenth century orthoepists seems to have upheld what is now the customary pronunciation; yet even here the authority of some of the greatest of modern poets has been occasionally cast in favor of the once accepted sound. In his poem of The Poet’s Mind, Tennyson, for instance, writes:

Holy water will I pour
Into every spicy flower.

The digraph is far from being limited to the sounds heard respectively in thou and you. Another one is that of long o, found, for illustration, in dough, soul, mould. There is still another sound—that of the so-called “broad a”—which is heard in brought, ought, and wrought. A fifth sound represented is that of the regular short u seen in would, could, and should. In cough and trough, as pronounced by many, there is a sixth sound represented. In the course of its travels through the vowel sounds the sign reaches that which we commonly call “short u.” There is no small number of words in which this pronunciation of it appears. Country, journey, trouble, flourish may be given as examples. Ou, in truth, has a remarkable record, not so much by the number of sounds it represents—in this it is approached by two or three other digraphs—but by the comparative largeness of the body of words in which several of these different sounds appear. In the latter respect, but not at all in the former, is it rivalled by the analogous ow. This, common as it is, has but two sounds. The first and most frequent is that heard in brown, down, and vowel; the second is the long o sound heard in such words as blow, grow, and below.

We now reach the digraphs of which the vowel u is the first letter. In a large number of words this has, if pronounced, the sound of w. Especially is this true of syllables upon which no accent falls, or at most a secondary accent. Nothing of this characteristic is seen in the case of uy—in which the diphthongal sound of i is heard in the two words buy and guy—but it is noticeable in the case of the first four vowels. We can see it illustrated by the ua of assuage, persuade, language; by the ue of conquest, request, and desuetude; by the ui of anguish, languish, cuirass; and by the uo of quote, quota, quorum. In this last case the u strictly belongs with q. Of ua, the first of these digraphs, all that needs to be said is that in certain words, such as guard and guardian, the u is not pronounced at all. The same statement can be made of ue in guess, guest, guerdon. It is as useless as it is silent. A plea has been put forth in justification of its existence on the theory that it acts as a sort of servile instrument to protect the hard sound of g. If this digraph were invariably so employed, it may be conceded that there would be some sense in its existence. But he who expects to find either sense or consistency in English orthography has strayed beyond the limits of justifiable ignorance. There is a large number of instances in which the consonant g continues to exhibit its hard sound when followed directly by e. Get and geese and gewgaw and eager and anger are a few of the words which could be adduced to show that there has never been felt any necessity of the presence of a protecting u to indicate this pronunciation.

When at the end of a word the digraph ue has often the sound of long u, as in blue, pursue, true, and rue. But no small number of instances occur in which it is entirely silent. This is especially noticeable in words derived from the Greek which have the final syllables logue or gogue. Catalogue, prologue, dialogue, demagogue, pedagogue, and synagogue will serve as examples. But the list of words in which this digraph is silent is far from being confined to those with these two terminations. Antique, oblique, intrigue, colleague, fatigue, rogue, and plague will testify to the uselessness of it as far as pronunciation is concerned, unless it be maintained that it justifies its existence by indicating that the preceding vowel has a long sound. If this be true, it ought not to appear when the vowel is short. One sees so much of the results of freak and wantonness in our spelling that it is permissible to cherish the fancy that any intelligent principle has been sometime somewhere at work in it, and that a feeling of this kind was the unconscious motive that led to the adoption of packet in place of pacquet and of lackey for lacquey; at any rate, of risk for the once prevalent risque and of check for cheque. But no such reason can be assigned for the ue of tongue. Its original was tunge. The final e ceased to be pronounced, and in course of time to be printed. The insertion of a u in the ending, after the fashion of the French langue, was an act of combined ignorance and folly.

The digraph ui follows in general the course of ue. As in the case of the latter the u was found unneeded in guess and guest, so it is equally unnecessary in guide and guile. Here again a not dissimilar sort of defence for it has been set up. Its retention, we are told, is desirable in order to indicate the diphthongal sound of i in these words. The argument is as futile as in the case of the preceding digraph. It illustrates forcibly the capabilities of our spelling in the way of confusing pronunciation that the same combination which is responsible for “long i” in guide and guile and disguise is equally responsible for the short i of guilt, guinea, and build. With the statement that ui has still another sound in such words as fruit, bruise, and recruit, we leave the consideration of the vowels and vowel sounds. But after the survey of the subject which has just been made, no one is likely to pretend that the pronunciation he hears of any one of these in a strange word will furnish him the least surety that he will be able to reproduce its authorized form in writing.

V
THE CONSONANTS

So much for the vowels. When we come to the consonants we are approaching much more solid phonetic ground. In a general way, they have remained faithful to the sounds they were created to indicate. Not but that here also there is need of reform. This will be made sufficiently manifest when details are given in the case of individual letters. But the disorganization of the consonant-system is slight compared with that of the vowel-system. There is, indeed, a fundamental difference between the two. With the vowels conformity to any phonetic law whatever is the exception and not the rule. With the consonants the reverse is the case. Fortunate it is for the English-speaking race that such is the fact. Were it otherwise, were there with the consonants the same degree of irregularity which exists with the vowels, the same degree of variableness in the representation of sounds, the same widely prevalent indifference to analogy, knowledge of English spelling would not be delayed, as it is now, for no more than two or three years beyond the normal time of its acquisition; it would be the work of a lifetime. Mastery of it, under existing conditions never fully gained by some, would in such circumstances never be acquired by anybody who learned anything else.

There is one pervading characteristic of the consonants which differentiates their position in the orthography from that of the vowels. Wherever they appear they have ordinarily the pronunciation which is theirs by right. Ordinarily, not invariably. There are exceptions that demand full discussion. Still, the usual way in which consonants vary from the phonetic standard is not by being pronounced differently but by not being pronounced at all. In some instances the useless letter represents the derivation; in others it defies it. They have been retained in the spelling, though never pronounced, either because they are found in the primitive from which they came; or they have been introduced into it under the influence of a false analogy, or as a consequence of a false derivation. In any reform of the orthography it may not be desirable in some cases to drop—at all events at the outset—these now silent letters. It assuredly would not be so wherever the tendency manifests itself to resume them in pronunciation.

There are four of the consonants which practically do not vary from phonetic law. They are never silent; they always indicate the precise pronunciation which they purport to indicate. In the case of two of them there is in each a single instance in which the rule does not hold good. In the preposition of, f has the sound of v. In the matter of inflection the temptation to retain this letter in spite of the change of sound has been successfully resisted. So we very properly say calves and wolves instead of calfs and wolfs, though this course exhibits what some must feel to be a scandalous tendency toward phonetic spelling. The other letter is m. The only exception to its regular pronunciation is found in the word sometimes spelled comptroller. Here it has the sound of n. But this has already been pointed out as a well-known spurious form based upon a spurious derivation. Its first syllable was supposed to come from the French compter and not from its real original, the Latin contra. The affection for this corrupt form now felt by some is in curious contrast with the attitude taken toward count both as a verb and a noun. These words were once often spelled like the corresponding French compte and compter. There was justification for this. They all came from the remote Latin original computare, in which the p is found. Naturally this particular spelling was especially prevalent in the sixteenth century, when derivation ran rampant in the orthography; but the practice extended much later. Had compt continued in use and fastened itself upon the language, we can imagine, but we cannot adequately express, the indignation that would now be felt by many worthy people at the proposal of any reformer to substitute for it count, and the picture of ruin to the speech that would be drawn as a result of such a wanton defiance of the derivation.

Let us now consider the unpronounced consonants. In the remote past such letters when no longer wanted were regularly dropped. Now they are as regularly retained. They are retained not because they are needed, but because they have become familiar to the eye. They naturally fall into three classes, according as they appear at the beginning, at the end, or in the middle of a word. To the first class belong g and k when followed by n; w followed by ho or by r; and the aspirate h. The failure to pronounce this last in certain words is too well known to need here more than a reference. Elsewhere, too, I have given an account of the gradual resumption of the sound of this letter.[23]

There are about half a dozen words in which an initial g is silent. Of these gnaw and gnat may be taken as examples. There are more than double this number in which an initial k before the same letter n is not heard. These are adequately represented, with the different vowels following, by knave, knee, knife, know, and knuckle. Still more frequently unsounded is an initial w. There are fully two dozen and a half of words in which this letter is not pronounced. The class finds satisfactory exemplification in who, whole, wrap, wrest, wrist, wrong, and wry. In making up these numbers it must be kept in mind that neither derivatives nor compounds are taken into account. Were such to be included, the list would be largely swelled.

In the cases just considered a letter once sounded has disappeared from the spoken tongue. The fact of its disappearance from pronunciation has not, however, induced men, as was once the practice, to discard it from the written tongue. But there are instances in which the initial consonant has never been heard at all in the utterance of any speakers. The words to which they belong are of foreign origin. They come to us with the foreign spelling. In many cases, or rather in most, they are from the Greek. The conspicuous examples are the c of czar, now frequently spelled tsar with the t sounded, the p of psalm and pseudo and of several compounds in which the psi of the Hellenic alphabet furnishes the initial letter. The same uselessness extends to ph—seen, for illustration, in the form phthisic—and to the p of words of Greek origin beginning with pt. It may be remarked in passing that there is a curious blunder in the spelling of the name of the bird called the ptarmigan. This is a pure Celtic word, which begins with t. To it a p was prefixed, possibly because it was supposed to be of Greek origin.

The final consonants which are retained in the spelling but are not heard in the pronunciation are b, n, h, t, w, and x. The words possessing them may be divided into two classes. In one the useless letter has a sort of claim to existence. It was there originally. Let us begin with the unpronounced final b. The native words ending in it are climb, comb, dumb, and lamb. They are common to the various Teutonic languages. In all of these they terminated originally with this consonant. To the list may be added plumb, ‘perpendicular,’ coming remotely from the Latin plumbum, ‘lead.’ The spelling of these words underwent the usual variations common before a fixed orthography had fastened itself upon the speech. Naturally the unpronounced b was not unfrequently dropped. This was especially true of climb and dumb. Take as an illustration Spenser’s line, where he speaks of a castle-wall,

That was so high as foe might not it clime.[24]

But after the reign of Elizabeth the useless letter gradually but firmly fixed its hold upon the spelling in the case of all these words. In this respect English has had a different development from that of other Teutonic tongues. Take modern German, for instance. For the word corresponding to climb it has replaced the original chlimban by klimmen; for chamb, ‘comb,’ it has substituted kamm; for dumb in Old High German tumb, it has dumm; for lamb, in Old High German lamb, it has lamm. The dropping of the final b seems to have wrought no observable harm to the language nor occasioned any grief—at all events, any present grief—to its users.

Still, it may be maintained in justification of the present spelling of these words that they are entitled to the final b on the ground of derivation. But no such plea can be put up in the case of those now to be considered. These are crumb, limb, numb, and thumb. In all of these the last letter is not only useless, but according to the term one chooses to employ, it is either a blunder or a corruption. It did not exist in the original. In truth, this unnecessary consonant threatened at one time to fasten itself also upon the name of the fruit called the plum. Especially was this noticeable in the best literature of the eighteenth century. An attack of common sense, to which the users of our orthography have been occasionally liable, prevented this particular word from carrying about the burden of the unpronounced b. In the case of most of the others it was not until the sixteenth century that the practice began of appending the unauthorized and unneeded letter. It took something of a struggle to foist it upon these words; but not so much, indeed, as will be required to loose the hold it has now gained over the hearts of thousands.

There are a few words, almost all of Latin derivation, in which a final n appears unsounded. Kiln is perhaps the only one of English extraction in which this peculiarity appears. In the case of most of them the retention of the letter may be defended—at least it may be palliated—on the ground that in the derivatives its pronunciation is resumed. In autumn, column, condemn, hymn, and limn the n is silent, but it gives distinct evidence of its existence in words like autumnal, columnar, condemnation, hymnal, limner, and solemnity. In fact, this resumption of the sound has at times been made to appear in other parts of the verbs containing this silent letter. Especially has this been true of hymning and limning, the participles of hymn and limn. It was a practice which much grieved certain of the earlier orthoepists. They took the ground that analogy forbade any sound not belonging to the principal verb itself to be heard in any of its parts. The observation is only noticeable for its revelation of the fact that it should enter into the head of any advocate of the existing orthography to set up analogy as a convincing reason for pronouncing any English word in a particular way.

Three of these final unpronounced letters do not need protracted consideration. In the digraph ow, ending such words as low, flow, and sow, the w serves no particular use. According to some it justifies its existence by indicating the quality of the preceding vowel. Its value in this respect may be estimated by comparing the pronunciation of bow, a missile weapon for discharging an arrow, with bow, an inclination of the head, or bow, the fore-end of a boat. The next letter t, when a final consonant, is invariably heard, save in some imperfectly naturalized words. Of these eclat and billet-doux may be taken as examples. In England, however—not in the United States—there is a single and singular survival of the original French pronunciation in the case of a word received into full citizenship. This is the noun trait, which came into the language in the eighteenth century. Naturally its final letter was at first not sounded. The tendency so to do, however, soon showed itself. Lexicographers authorized it, indeed favored it; but for some inexplicable reason Englishmen have never taken kindly to the complete naturalization of the word. “The t,” said Walker, at the end of the eighteenth century, “begins to be pronounced.” Had he been living at the end of the nineteenth, he would have been justified in saying precisely the same thing as regards England. It was beginning then; it is beginning now; but it is only beginning.

A final h is not pronounced when preceded by a vowel; when preceded by the consonant g it forms a digraph which will be considered later. There are fewer than a dozen words of the former class in which it appears. Among these are the interjections, ah, eh, and oh. Here again, as in the case of w, the existence of the letter is defended on the ground that it indicates the quality of the preceding vowel. Yet for this purpose it can hardly be deemed a necessity. We use it in the case of ah; but we get along very well without it in the case of ha. This, too, was formerly sometimes spelled hah. Oh, likewise, was once widely found in the very instances and the very senses where we now use the single letter O. In two other words, Messiah and hallelujah, the h may be retained because of the sacredness of associations which have gathered about them. Yet the former word was itself a sixteenth-century alteration of the previous Messias.

The unpronounced final k belongs strictly to the class of double letters of which it is not my purpose to treat. It invariably follows c, and is really nothing but a duplicate of it. Still, as the sign is a different representation of the same sound, it may be well to bestow upon it a brief attention. At the end of the eighteenth and the beginning of the nineteenth century it was dropped, after a warm contest, from words of Latin derivation. But the reform did not extend to those of native origin. In many cases a k has been added to words which originally ended in c. Especially was this true of monosyllables. Thus the earliest form of back was bæc, of sack was sac, of sick was seoc. This was the case not only with a good many monosyllables to which a k is now appended, but to a certain extent with dissyllables also. The fact is best exemplified in the words which have the ending ock. This sometimes represents the early English diminutive uc, which became later oc or ok. From the point of view of derivation the modern spelling is distinctly improper. Thus, for illustration, bullock, haddock, hassock, hillock, and mattock were in their earliest known forms bulluc, haddoc, hassuc, hilloc, and mattuc. Several words not of native origin have also adopted this ending. Hammock, from the Spanish hamaca, itself of Carib origin, has conformed to it. It has supplied itself with a final k. During the last century havoc managed to get rid of this consonant, with which it had been encumbered, without exciting any special remark. But now that the uselessness of a letter has become to many one of the chief recommendations of the spelling, the dropping of an unnecessary k from any of the other words of this class would bring unspeakable anguish to thousands.

There are more consonants which are unpronounced in the middle of words than at the beginning or the end. They are b, c, l, g, h, p, s, t, w, and z. In the case of some of them—the two last, for instance—the words in which the unpronounced letter appears are very few. In rendezvous z is not sounded. It is the only instance in which this consonant is not heard, and this is due to the fact that it is not heard in its French original. Again, it is only in answer, sword, and two that the medial w is silent. Unpronounced consonants are more frequent in the case of the other letters, but, after all, they are not numerous in themselves. Still, their presence has its usual effect. In every instance it raises a stumbling-block in the way of the proper pronunciation. Furthermore, it has in some cases either hidden the right derivation entirely or given a wrong idea of it.

Take the example of the medial b in debt and doubt. These words, coming originally from the French, were introduced into the language with the spelling, dette, det, and doute, dout. So for a long time they were spelled. Deference to the remote Latin original, which sprang up with the revival of learning, introduced the unauthorized b into the world. It has already been pointed out that this has given an opportunity, which has been fully improved, for the devotees of derivation to exhibit their usual inconsistency. When the presence of unpronounced letters in the case of other words presents an obstacle to correct pronunciation, then its retention is insisted upon as essential to our knowledge of its immediate origin, to the purity of the language itself, and to the happiness of those speaking it. But no advocate of the existing orthography could be induced to part with the b of debt and doubt, though its presence comes into direct conflict with the views he is championing in the case of other words. At times attempts were apparently made to pronounce the inserted b. In the full Latinized form debit, which was early in use, there was no difficulty. Indeed, it was a necessity. Not so in the form debt. Yet it is evident from Love’s Labour’s Lost that there were men who sought to accomplish this feat. It is difficult to ascertain whether speaker or hearer suffered more in consequence of this effort. If unsuccessful, it was the speaker; if successful, the pain was transferred to the hearer.

One curious blunder has been foisted upon our spelling by the desire of men to go back to the Latin original of these words instead of contenting themselves with the immediate French one. The insertion of b is bad enough in redoubted and redoubtable. These came to us from the latter tongue, and at first appeared in English in the forms redouted and redoutable. Later the classical influence made itself felt and the b was inserted. Palliation for it could be pleaded on the ground that the letter belonged to the remote original. But no defence of this sort could be of avail in the case of the word denoting the military outwork called a redoubt. This has not the slightest connection either in sense or origin with the two adjectives just specified. It comes directly from the French redout and remotely from the Latin reductus, ‘withdrawn,’ ‘retired,’ which received at a later period the meaning of ‘a place of refuge.’ But it was ignorantly supposed that it came from the same source as the verb redoubt and its past participle redoubted. So from the beginning of the introduction of the word into the language, in the early part of the seventeenth century, an unauthorized b was made part of it. It is now dear to the hearts of millions. What the blundering of one age perpetrated the superstition of succeeding ages has invested with peculiar sanctity.

The cases in which c follows s present several choice examples of the vagaries which make English orthography a wonder to those who study its history, and a perpetual joy and boast to those who in this matter succeed in keeping the purity of their ignorance from being denied by the slightest stain of knowledge. In the words scene, scepter, and sciatica, coming directly or remotely from the Greek, the letter represents an original k. So, useless as it is, its retention may be defended on the ground that if it be not the same letter, it ought to be, since it has the same value. The similar apology of respect for derivation may be urged for the unpronounced c of science, scintilla, and sciolist. But in the case of scent, scion, scimitar, scissors, and scythe, no such plea can be made. In the instance of all these there is not the slightest justification for the unnecessary c. Scent comes from the Latin sent-ire, ‘to perceive.’ Until the seventeenth century it was regularly spelled sent. Scythe, from the Anglo-Saxon sîthe, once frequently and now occasionally has its strictly correct etymological form. Scion is from the old French sion. Scimitar and scissors have had a wide variety of spellings during the course of their history. English orthography has exhibited, as is not unusual, a perverse preference for the ones which depart furthest from the pronunciation.

The instances where g is silent within a word are those in which it is found preceding m or n. Its presence it owes, in most instances, to derivation. Examples of it can be found in a number of words of Greek extraction, of which paradigm, diaphragm, and phlegm may be given. With a following n it can be represented by campaign, feign, sign, and impugn. As has been the case with the final n of certain words, so also the pronunciation of the g is resumed in the derivatives. That may be deemed by some a justification for its retention in the primitive—at least, for the time being. With sign we have signify, with malign we have malignity, with phlegm we have phlegmatic. But the g is a particularly ridiculous intruder in the words foreign and sovereign. The former is from the Old French forein, which itself comes from the popular Latin foraneus, and this in turn comes from the classical Latin foras, ‘out of doors.’ Sovereign is a spelling just as bad. It comes from the Old French sovrain, the Low Latin superanus, ‘supreme,’ which was formed upon the preposition super, ‘above.’ The insertion of a g was a blunder for which our race has the sole responsibility.

There are two kinds of words in which h is silent following an initial letter. This is invariably true of words of Greek extraction beginning with rh. Rhetoric, rheumatic, and rhubarb may serve as specimens. In these, as in those like them, the h was wanting in Old French. Consequently, it was at first wanting in English also. But the deference to derivation which prevailed among the classically educated after the revival of learning, raised havoc here with the spelling as it did in so many other instances. The unpronounced h was inserted into all these words. This began in the sixteenth century. It gradually established itself firmly in the orthography. There it has remained ever since, though no one pretends that it serves any purpose save that of indicating to the few, who do not need to be informed, that the aspirate existed in the original from which these words were derived. But even this pitiable reason cannot be pleaded in the case of the noticeable words in which h follows an initial g. These are ghastly and aghast, ghost, and gherkin. In not one of them, except the last, did h appear till many hundred years after the words had been in existence. To not one of them does the useless letter belong by right. Indeed, it was apparently not till the nineteenth century that it was foisted into gherkin as the regular spelling, though it had cropped up before. There would be just as much sense in spelling German as Gherman, and goat as ghoat, as there is in the intrusion of the h into the words just mentioned. This is equally true of anchor.

There is, however, one further peculiarity about this letter. In the spelling of certain words it follows w, in the pronunciation of them it precedes it. But the fashion of suppressing the sound of the aspirate in the combination wh is very characteristic of the speech of England, at least of some parts of it. The prevalence of this sort of pronunciation which makes no distinction, for example, between where and wear, between Whig and wig, between while and wile, was a subject of great, and it may be added, of justifiable grief to the earlier orthoepists. Walker complained bitterly of the extent of its use in London. He was anxious that men should “avoid this feeble Cockney pronunciation which is so disagreeable to a correct ear.” Fortunately for the speech this suppression of the aspirate has not extended much beyond the southern half of England. In America it rarely takes place. There is, therefore, every likelihood of this pronunciation being eventually crushed, not so much because of its own inherent viciousness as by the mere weight of numbers.

There is a limited body of words in which l and p are silent. The former letter in such cases as balm and calm, for instance, may perhaps have been effective in preserving the sound of the preceding vowel. The most signal example of its appearance, where it has no justification for its existence, is in the word could. This takes the place of the earlier and more correct coude, coud. The l was introduced by a false analogy with would and should. These two last words, it may be added, at times dropped this letter, to which etymologically they were entitled, out of deference to the pronunciation, just as could, though not entitled to it, assumed it in defiance of the pronunciation.

The most noticeable instances in which p is not pronounced are when it follows m and is itself followed by t. Empty, tempt, prompt, and sumptuous will supply a sufficient number of illustrations. In most of these cases the letter still appears because it was in the original. In empty, however, it is a later insertion. There are two or three sporadic instances in which a p is present but fails to be called upon for duty. Such are raspberry and receipt. In the first of these two rasberry seems to have been for a long time the preferred spelling. Unless there is a prospect that the sound of the letter will be resumed in the pronunciation, there is no apparent reason why we should not go back to the once more common form. But receipt, with the allied conceit and deceit, furnishes as good an illustration as can well be offered of the vagaries of English orthography, and of the system which has prevailed in and the sense which has presided over its development. These three words all come remotely from the three closely allied participial forms receptus, conceptus, and deceptus. The earlier most common spelling of the first was receit or receyt. While the form with the inserted p existed previously, it was not till the Elizabethan period that it began to be much in evidence. Furthermore, it was not till the latter part of the seventeenth century that the unnecessary letter established itself in the unfortunate word. Conceit and deceit went through what was in many respects the same experience. The forms conceipt and deceipt were found not unfrequently. But in them the p failed to maintain itself. So words from a common Latin root have developed two different ways of spelling, with not the slightest reason in the nature of things why any distinction whatever should be made between them.

The silence of s in some few words, such as isle, aisle, and island, has already been mentioned. In viscount it is also suppressed, doubtless in deference to the French original. But in the middle of words t is far more frequently left unpronounced than s. This is especially noticeable when it is followed by le on the one hand, as can be seen in castle, wrestle, thistle, ostler, and rustle; on the other hand, when followed by en, as in fasten, hasten, listen, and moisten. There are a few other words besides those with these endings in which it is silent. Such are Christmas, chestnut, mortgage, bankruptcy. That it should not be heard in words of French origin like billet-doux and hautbois is not hard to understand; they have never been fully naturalized.

This exhausts the list of simple consonants that are found in the written language, but are not heard in the spoken. There remains, however, a digraph which is encountered too frequently not to receive brief mention. This is gh, both at the end and in the middle of words. In these positions it once stood for something. It had, therefore, originally a right to the place in which it now appears. But the guttural sound it indicated disappeared long ago from the usage of all of us. Even the knowledge that it had ever existed has disappeared from the memory of most of us, if it was ever found there. Accordingly it serves now no other purpose than to act as a sort of tombstone to mark the place where lie the unsightly remains of a dead and forgotten pronunciation. The useless digraph is still seen at the end of numerous words of which weigh, high, and dough may be taken as examples. Again an unpronounced medial gh is seen in neighbor and a large number of words ending in ght, such as caught, height, fight, and thought. In many of these words the digraph was frequently dropped in those earlier days when there was a perverse propensity to make the spelling show some respect to the pronunciation. High, for instance, often appeared in the forms hye, hy; nigh in the forms nye, ny. This is now all changed. The disposition to pander to any sneaking desire to bring about a scandalous conformity between orthography and orthoepy is steadily frowned upon by those who have been good enough to take upon their shoulders the burden of preserving what they are pleased to call the purity of the English language.

This survey of the subject, brief as it is, brings out distinctly the superiority of the consonant system over the vowel, in the matter of unpronounced letters. Far from perfect as is the former, it shines by contrast with the latter. The useless consonant appears in but a few words, where the useless vowel appears in scores. But when we pass on to the cases in which the sign is represented by any but its legitimate sound, the contrast between the two classes of letters becomes far more noticeable. It is the superiority in this particular which alone makes our present spelling endurable. Most of the consonants, if pronounced at all, have in all cases one and the same sound. Any possible acquisition of the speech in the term of a man’s natural life has depended upon the fact that these members of the alphabet are in general really phonetic. Their faithfulness to their legitimate sounds stands in sharpest contrast to the almost hopeless disorganization which has overtaken the vowels. In the case of some of the consonants there is never any variation from their proper pronunciation. In the case of others the exceptions to the regular practice are purely sporadic. The p of cupboard, for instance, has the sound of b, the j of hallelujah has the sound of y. Even these exceptions which have prevailed in the past there has been a tendency to reduce, owing to the operation of agencies of which there will be occasion to speak later.

This last statement needs modification in the case of one letter. In modern times there has been a tendency to represent the sound of t in the preterite and past participle by d, or, rather, ed. As compared with the usage of the past, this practice has made a good deal of headway. It is the substitution of a formal regularity of spelling which appeals to the eye over its proper use to indicate the sound to the ear. We have not yet got so far as to write sleeped for slept or feeled for felt, but we have frequently dwelled for dwelt and builded for built. This is all proper enough if the d sound is given to the ending by pronouncing the word, as is often done, as a dissylable. But no reason can be pleaded for it if t is heard as the termination. In this matter we are far behind our fathers.

Take the usage of Spenser, as illustrated on this point in the first canto of the first book of the Faerie Queene. This contains about five hundred lines. In every case whenever a preterite or past participle has the sound of t, it is spelled with t. In this one canto—and it fairly represents all the others—can be found the preterites advaunst, approcht, chaunst, enhaunst, forst, glaunst, grypt, knockt, lept, lookt, nurst, pusht, y-rockt, stopt, and tost. Along with these are to be seen as past participles accurst, enforst, mixt, past, promist, stretcht, vanquisht, and wrapt. Now, to a certain extent this is an unfair illustration. No one can read the Faerie Queene without becoming aware that Spenser was a good deal of a spelling reformer. Necessarily, he was largely dominated by the ignoble idea that orthography should have a close connection with the pronunciation. Still, though in certain particulars he took very advanced ground, he only practiced on a large scale what on a small scale was followed by very many of his contemporaries and immediate successors.

We pass on now to the consideration of the six sounds for which the alphabet has no special sign whatever. Two of them are the surd and sonant sounds, already considered, for which the digraph th has become the common representative. It may be right to add that this same digraph is also equivalent in a few cases to the simple t, as in thyme and Thames. The four other sounds can be recognized perhaps most easily in the ch of church, the ng of bring, the sh of ship, and the s of pleasure. But here, as elsewhere in our orthography, reigns the usual lawlessness. The signs here given represent other sounds than those just specified. Take the case of ng. Any one can detect at once the difference in the pronunciation of this digraph by contrasting it as heard in singer and as heard in finger. Nor has ch been limited to the sound indicated in chair, cheer, child, choose, and churn. It has another, perhaps more frequently denoted by sh in the beginning, middle, and end of words, as, for illustration, in chaise, machine, and bench. It has likewise the sound of k in many words, especially in those of Greek origin, such as character, mechanic, monarch. The uncertainty caused by this variety of pronunciation is particularly noticeable in words in which arch appears as the initial syllable. In archangel, for instance, ch has one pronunciation, in archbishop it has another. The difference between the two must therefore be painfully learned. There is, furthermore, the sporadic example of choir, in which ch has the sound of kw, ordinarily represented by qu. But choir was a late seventeenth-century importation into the language. Though to some extent it has replaced the original form quire, it has invariably retained the pronunciation of that word.

Finally, there are the two sounds specified above, as denoted by the s of pleasure and the sh of ship. The former has a respectable number of signs to indicate it. Besides the s found in such words as measure, usury, enclosure, it is represented by si, as seen in decision, evasion, occasion; by z, as in azure, razure, seizure; by zi, as in glazier, grazier, vizier. It is, however, the second of these sounds that has the greatest variety of signs to denote it. In this respect it rivals many of the vowels or vowel combinations, and surpasses some of them. It is heard in the ce of ocean, and in particular in no small number of words mainly scientific, with the ending aceous, such as cretaceous and cetaceous; in the ci of words like social, gracious, suspicion; in the s of sure, sugar, censure, nauseate; in the t of satiate, expatiate, substantiate; in the ti of martial, patient, nation, and the vast number of words which have the termination tion; in xi in anxious, obnoxious, complexion; in sci in conscience, prescience; in si, as seen in no small number of words, such as mansion, vision, explosion. Finally, to illustrate the confusion which in the case of these signs has been still further confounded, we may instance the ci of social with the pronunciation just indicated, and the ci of the related word society with a pronunciation entirely different. A precisely similar observation could be made of ti in the case of the words satiate and satiety.

Enough has certainly now been said to put beyond question the fact of the irrepressible conflict which goes on in our language between orthography and orthoepy, and to make clear its nature. The treatment of the subject has, indeed, been far from complete. Nothing whatever has been said on the large subject of the representation of sounds in the unaccented syllables. No account has been given of the usage of some of the letters or combinations of letters. In particular, in the matter of doubling the letters both in accented and unaccented syllables, contradictions and incongruities abound with us on a scale which ought to bring peculiar happiness to those devotees of the present orthography who believe that the worse a language is spelled the more distinctly it is to its credit. Still, of this characteristic there has been no consideration. Furthermore, page after page could have been taken up with illustrative examples of the anarchy of all sorts which reigns in every nook and corner of our spelling. We write, for instance, knowledge with a d; but the place with the same terminating syllable where we go presumably to acquire it, which we call a college, we are careful to write without a d. In the past one finds at times the forms knowlege and colledge. It is nothing but an accident of usage that we are not employing them now instead of the ones we have adopted.

It would be easy to go on multiplying examples of these inconsistencies. But though all that could be said is far from having been said, surely enough has been given to prove beyond possibility of denial the existence of the chaotic condition which prevails. Furthermore, while the subject has been by no means exhausted, the same statement cannot safely be made of the patience of the reader, to say nothing of that of the writer. If any one of the former body finds it tedious to wade through the account of the situation which has been given in the preceding pages, let him bear in mind how much more tedious it was for the author to prepare it. If he finds it exceedingly tedious, let him take to himself a sort of consolation in the reflection of how easily it could have been made even more so. Instead, therefore, of complaining of the abundance of minute detail which I have supplied, he ought to be thankful to me for keeping back so much of it as I have done. Moreover, as Heine pointed out long ago, the reader has at his command a resource to which he can always betake himself when his powers of endurance give out. He can skip. This is a blessed privilege denied to the writer.

Incomplete, however, as has been the survey of the subject, it has been sufficient to give a fairly satisfactory idea of the way in which the orthography represents, or rather misrepresents, the pronunciation. It makes manifest beyond dispute the truth of the intimation conveyed at the outset that the form of a particular word is often, with us, little more than a fortuitous concourse of unrelated letters in which neither they nor the combinations into which they enter can be relied upon to indicate any particular sound. In addition, hundreds of those which appear in the spelling have no office in the pronunciation. Genuine derivation has led to the retention of some, spurious derivation to the introduction of others. There are, consequently, few of the common words of our language which cannot be spelled with perfect propriety in different ways, sometimes in half a dozen different ways, if the analogy be followed of words similarly formed and pronounced. Our orthography is, therefore, often a matter of contention and always a matter of study. Knowledge of the accepted form of words must be gained in each case independently, for there exists no general principle, the observance of which will guide the learner to a correct conclusion.

As an inevitable result, the acquisition of spelling never calls into exercise, with us, the reasoning faculties. On the contrary, its direct effect is to keep them in abeyance. The ability to spell properly is an intellectual act only to the extent that attention and recollection are intellectual acts. It can and not unfrequently does characterize persons who are very far from being gifted with much mental power. All who attain proficiency in it are compelled to spend time which, under proper conditions, could have been far more profitably employed. There are men who do not attain it at an early age, and some even who never attain it at all. Moore, for illustration, speaking of Byron, tells us that spelling was “a very late accomplishment with him.”[25] The case of William Morris was far worse. This poet never learned to spell at all. The fact is recorded by his biographer. In speaking of the beauty of his handwriting, he had to admit the failure of his orthography to reach the standard set by it. “The subsidiary art of spelling,” he writes, “was always one in which he was liable to make curious lapses. ‘I remember,’ the poet once said, ‘being taught to spell and standing on a chair with my shoes off because I made so many mistakes.’ In later years several sheets of The Life and Death of Jason had to be cancelled and reprinted because of a mistake in the spelling of a perfectly common English word; a word, indeed, so common that the printer’s reader had left it as it was in the manuscript, thinking that Morris’ spelling must be an intentional peculiarity.”[26]

The ignorance which exists in regard to the orthographic situation is bad enough; but the superstition which has been born of it is still worse. It is assumed to have come down to us pure and perfect from a remote past. Hence, it must be religiously preserved in all its assumed sacredness and genuine uncouthness. Even improvements which could be made with little difficulty, which would have no other result than bringing about with the least possible friction uniformity in certain classes of words—these slight alterations are assailed with almost as much earnestness and virulence as would be encountered by sweeping changes designed to make the spelling really phonetic.

As men are more apt to be interested in particular illustrations than in general discussion, it may be worth while to follow up the survey of the situation which has just been given with an account in detail of the history of a special class of words. In this once prevailed the tendency to bring about absolute uniformity. The movement was arrested before the desired result was attained. It left a few over thirty examples as exceptions to the general practice. In the derivatives of some of these it went back to the regular rule and consequently contributed exceptions to the exceptions. This condition of things has endeared these anomalies to the hearts of thousands. The class itself consists of the words ending in or or our. About the proper way of spelling this termination controversy has raged for more than a hundred years. The examination of the whole class can be best carried on by selecting one of the words belonging to it as typical of all. To its story the next chapter will be largely confined.