The English Language

§ 190. In the way of orthography, the most characteristic difference between the English and Scotch is the use, on the part of the latter, of qu for wh; as quhen, quhare, quhat, for when, where, what. The substitution of sch for sh (as scho for she), and of z for the Old English ȝ (as zour for ȝeowr, your), is as much northern English as Scotch.

In pronunciation, the substitution of d for ð (if not a point of spelling), as in fader for father; of a for o, as báith for both; of s for sh, as sall for shall; and the use of the guttural sound of ch, as in loch, nocht, are the same.

The ejection of the n before t, or an allied sound, and the lengthening of the preceding vowel, by way of compensation, as in begouth for beginneth, seems truly Scotch. It is the same change that in Greek turns the radical syllable ὀδοντ into ὀδούς.

The formation of the plural of verbs in -s, rather than in -th (the Anglo-Saxon form), is Northern English as well as Scotch:—Scotch, slepys, lovys; Northern English, slepis, lovis; Old English, slepen, loven; Anglo-Saxon slepiað, lufiað.

The formation of the plural number of the genitive case by the addition of the syllable -is (blastis, birdis, bloomis), instead of the letter -s (blasts, birds, blooms), carries with it a metrical advantage, inasmuch as it gives a greater number of double rhymes.

The same may be said of the participial forms, affrayit, assurit, for affrayd, assured.

Concerning the comparative rate of change in the two languages no general assertion can be made. In the Scotch words sterand, slepand, &c., for steering, sleeping, the form is antiquated, and Anglo-Saxon rather than English. It is not so, however, with the words thai (they), thaim (them), thair (their), compared with the contemporary words in English, heo, hem, heora. In these it is the Scottish that is least, and the English that is most Anglo-Saxon.

CHAPTER IV.

OF CERTAIN UNDETERMINED AND FICTITIOUS LANGUAGES OF GREAT BRITAIN.

§ 191. The languages mentioned in the present chapter claim their place on one ground only,—they have been the subject of controversy. The notice of them will be brief. The current texts upon which the controversies have turned will be quoted; whilst the opinion of the present writer is left to be collected from the title of the chapter.

The Belgæ.—By some these are considered a Germanic rather than a Celtic tribe; the view being supported by the following extracts from Cæsar:—"Gallia est omnis divisa in tres partes; quarum unam incolunt Belgæ, aliam Aquitani, tertiam, qui ipsorum lingua Celtæ, nostra Galli, appellantur. Hi omnes lingua, institutis, legibus inter se differunt. Gallos—a Belgis Matrona et Sequana dividit."—B. G. i. "Belgæ ab extremis Galliæ finibus oriuntur."—B. G. ii. "Quum ab his quæreret, quæ civitates, quantæque in armis essent, et quid in bello possent, sic reperiebat: plerosque Belgas esse ortos a Germanis, Rhenumque antiquitùs transductos, propter loci fertilitatem ibi consedisse; Gallosque, qui ea loca incolerent, expulisse; solosque esse qui patrum nostrorum memoria, omni Gallia vexata Teutones Cimbrosque intra fines suos ingredi prohibuerunt."—B. G. ii. 4. "Britanniæ pars interior ab iis incolitur quos natos in insulâ ipsâ memoriâ proditum dicunt: maritima pars ab iis, qui prædæ ac belli inferendi causa ex Belgio transierant."—B. G. v. 12.

§ 192. The possibly Germanic origin of the Belgæ, and the Belgic element of the British population, are matters which bear upon the question indicated in § 10, or that of the Germanic influences anterior to A.D. 449.

They have a still more important bearing, the historian over and above identifying the Belgæ with the Germans, affirms that what applies to the Belgæ applies to the Picts also.

Now this is one of the arguments in favour of the doctrine exhibited (and objected to) in pp. 124-127, and the extent of questions upon which it bears, may be collected from the following quotation:—"A variety of other considerations might be mentioned, which, although they do not singly amount to proof, yet merit attention, as viewed in connexion with what has been already stated.

"As so great a part of the eastern coast of what is now called England was so early peopled by the Belgæ, it is hardly conceivable that neither so enterprising a people, nor any of their kindred tribes, should ever think of extending their descents a little farther eastward. For that the Belgæ and the inhabitants of the countries bordering on the Baltic, had a common origin, there seems to be little reason to doubt. The Dutch assert that their progenitors were Scandinavians, who, about a century before the common era, left Jutland and the neighbouring territories, in quest of new habitations.^[29] The Saxons must be viewed as a branch from the same stock; for they also proceeded from modern Jutland and its vicinity. Now, there is nothing repugnant to reason in supposing that some of these tribes should pass over directly to the coast of Scotland opposite to them, even before the Christian era. For Mr. Whitaker admits that the Saxons, whom he strangely makes a Gaulic people, in the second century applied themselves to navigation, and soon became formidable to the Romans.^[30] Before they could become formidable to so powerful a people, they must have been at least so well acquainted with navigation as to account it no great enterprise to cross from the shores of the Baltic over to Scotland, especially if they took the islands of Shetland and Orkney in their way.

"As we have seen that, according to Ptolemy, there were, in his time, different tribes of Belgæ, settled on the northern extremity of our country: the most natural idea undoubtedly is, that they came directly from the Continent. For had these Belgæ crossed the English Channel, according to the common progress of barbarous nations, it is scarcely supposable that this island would have been settled to its utmost extremity so early as the age of Agricola.

"There is every reason to believe, that the Belgic tribes in Caledonia, described by Ptolemy, were Picts. For as the Belgæ, Picts, and Saxons seem to have had a common origin, it is not worth while to differ about names. These frequently arise from causes so trivial, that their origin becomes totally inscrutable to succeeding ages. The Angles, although only one tribe, have accidentally given their name to the country which they invaded, and to all the descendants of the Saxons and Belgæ, who were by far more numerous.

"It is universally admitted, that there is a certain national character, of an external kind, which distinguishes one people from another. This is often so strong that those who have travelled through various countries, or have accurately marked the diversities of this character, will scarcely be deceived even as to a straggling individual. Tacitus long ago remarked the striking resemblance between the Germans and Caledonians. Every stranger, at this day, observes the great difference and complexion between the Highlanders and Lowlanders. No intelligent person in England is in danger of confounding the Welsh with the posterity of the Saxons. Now, if the Lowland Scots be not a Gothic race, but in fact the descendants of the ancient British, they must be supposed to retain some national resemblance of the Welsh. But will any impartial observer venture to assert, that in feature, complexion, or form, there is any such similarity as to induce the slightest apprehension that they have been originally the same people?"^[31]

It is doubtful, however, whether Cæsar meant to say more than that over above certain differences which distinguished the Belgæ from the other inhabitants of the common country Gallia, there was an intermixture of Germans.

The import of a possibly Germanic origin for the Belgæ gives us the import of a possibly Germanic origin for—

§ 193. The Caledonians.—A speculative sentence of Tacitus indicates the chance of the Caledonians being Germanic:—"Britanniam qui mortales initio coluerint, indigenæ an advecti, ut inter barbaros, parum compertum. Habitus corporum varii: atque ex eo argumenta: namque rutilæ Caledoniam habitantium comæ, magni artus, Germanicam originem adseverant."—Agricola, xi.

The continuation of the passage quoted in § 193 has induced the notion that there have been in Britain Spanish, Iberic, or Basque tribes:—"Silurum colorati vultus, et torti plerumque crines, et posita contra Hispania, Iberos veteres trajecisse, easque sedes occupâsse fidem faciunt."—Agricola, xi.

As this, although an opinion connected with the history of the languages of Great Britain, is not an opinion connected with the history of the English language, it is a question for the Celtic, rather than the Gothic, philologist. The same applies to the points noticed in §§ 136-138. Nevertheless they are necessary for the purposes of minute philological analysis.

§ 194. As early as the year A.D. 1676, an opinion was advanced by^[32] Aylett Sammes, in a work entitled Britannia Antiqua Illustrata, that the first colonisers of Ireland were the merchants of Tyre and Sidon. In confirmation of this opinion the existence of several Eastern customs in Ireland was adduced by subsequent antiquarians. Further marks of an Eastern origin of the Irish were soon found in the Gaelic dialect of that country. Finally, the matter (in the eyes at least of the national writers) was satisfactorily settled by the famous discovery, attributed to General Vallancey, of the true meaning of the Carthaginian lines in Plautus.

In the Little Carthaginian (Pœnulus) of the Latin comic writer Plautus, a portion of the dialogue is carried on in the language of Carthage.

That the Punic language of Carthage should closely resemble that of the mother-city Tyre, which was Phœnician; and that the Phœnician of Tyre should be allied to the language of Palestine and Syria, was soon remarked by the classical commentators of the time. Joseph Scaliger asserted that the Punic of the Pœnulus differed but little from pure Hebrew—"Ab Hebraismi puritate parum abesse."

Emendated and interpreted by Bochart, the first ten lines of a speech in Act v. s. 1. stand thus:—

Six lines following these were determined to be Liby-Phœnician, or the language of the native Africans in the neighbourhood of Carthage, mixed with Punic. These, it was stated, had the same meaning with the ten lines in Carthaginian.

The following lines of Plautus have, by all commentators, been viewed in the same light, viz. as the Latin version of the speech of the Carthaginian.

Guided by the metrical paraphrase of the original author, Bochart laid before the scholars of his time a Latin version, of which the following is an English translation:—

Close Translation of Bochart's Latin Version.

To professed classics and to professed orientalists, the version of Bochart has, on the whole, appeared satisfactory. Divisions of opinion there have been, it is true, even amongst those who received it; but merely upon matters of detail. Some have held that the Punic is Syriac rather than Hebraic, whilst others have called in to its interpretation the Arabic, the Maltese, or the Chaldee; all (be it observed) languages akin to the Hebrew. Those who look further than this for their affinities, Gesenius^[33] dismisses in the following cavalier and cursory manner:—"Ne eorum somnia memorem, qui e Vasconum et Hiberniæ linguis huic causæ succurri posse opinati sunt; de quibus copiosius referre piget."

The remark of Gesenius concerning the pretended affinities between the Punic and Hibernian arose from the discovery attributed to General Vallancey; viz. that the speech in Plautus was Irish Gaelic, and consequently that the Irish was Carthaginian, and vice versâ. The word attributed is used because the true originator of the hypothesis was not Vallancey, but O'Neachtan.

The Gaelic Version.

In English.

From the quotations already given, the general reader may see that both the text and the translation of Plautus are least violated in the reading and rendering of Bochart, a reading and rendering which no Gothic or Semitic scholar has ever set aside.

§ 195. The hypothesis of an aboriginal Finnic population in Britain and elsewhere.—A Celtic population of Britain preceded the Germanic. Are there any reasons for believing that any older population preceded the Celtic?

The reasoning upon this point is preeminently that of the Scandinavian (i.e. Danish, Swedish, and Norwegian) school of philology and ethnology.

Arndt, I believe, was the first who argued that if the so-called Indo-European nations were as closely connected with each other as they are generally considered, their separation from the common stock must have been subsequent to the occupation of Europe by some portion or other of the human species—in other words, that this earlier population must have been spread over those areas of which the Indo-Europeans took possession only at a later period.

That the divisions of such an earlier population were, at least, as closely connected with each other as the different members of the so-called Indo-European class, was a reasonable opinion. It was even reasonable to suppose that they were more closely connected; since the date of their diffusion must have been nearer the time of the original dispersion of mankind.

If so, all Europe (the British Isles included) might have had as its aborigines a family older than the oldest members of the Indo-European stock; a family of which every member may now be extinct, or a family of which remains may still survive.

Where are such remains to be sought? In two sorts of localities—

1. Parts beyond the limits of the area occupied by the so-called Indo-Europeans.

2. Parts within the limits of the so-called Indo-Europeans; but so fortified by nature as to have been the stronghold of a retiring population.

What are the chief parts coming under the first of these conditions?

a. The countries beyond the Indo-Europeans of the Scandinavian and Slavonic areas, i.e. the countries of the Laplanders and Finnlanders.

b. The countries beyond the Indo-Europeans of the Iranian stock, i.e. the Dekkan, or the country of those natives of India (whatever they may be) whose languages are not derived from the Sanscrit.

What are parts coming under the second of these conditions?

a. The Basque districts of the Pyrenees, where the language represents that of the aborigines of Spain anterior to the conquest of the Roman.

b. The Albanians.—Such the doctrine of the continuity of an ante-Indo-European population, from Cape Comorin to Lapland, and from Lapland to the Pyrenees. There is some philological evidence of this: whether there is enough is another matter.

This view, which on its philological side has been taken up by Rask, Kayser, and the chief Scandinavian scholars, and which, whether right or wrong, is the idea of a bold and comprehensive mind, as well as a powerful instrument of criticism in the way of a provisional theory, has also been adopted on its physiological side by the chief Scandinavian anatomists and palæontologists—Retzius, Eschricht, Niilson, and others. Skulls differing in shape from the Celtic skulls of Gaul, and from the Gothic skulls of Germany and Scandinavia, have been found in considerable numbers; and generally in burial-places of an apparently greater antiquity than those which contain typical Celtic, or typical Gothic crania. Hence there is some anatomical as well as philological evidence: whether there is enough is another question.

PART III.

SOUNDS, LETTERS, PRONUNCIATION, SPELLING.

————

CHAPTER I.

GENERAL NATURE OF ARTICULATE SOUNDS.

§ 196. To two points connected with the subject of the following Chapter, the attention of the reader is requested.

I. In the comparison of sounds the ear is liable to be misled by the eye.

The syllables ka and ga are similar syllables. The vowel is in each the same, and the consonant is but slightly different. Now the words ka and ga are more allied to each other than the words ka and ba, ka and ta, &c., because the consonantal sounds of k and g are more allied than the consonantal sounds of k and b, k and t.

Comparing the syllables ga and ka, we see the affinity between the sounds, and we see it at the first glance. It lies on the surface, and strikes the ear at once.

It is, however, very evident that ways might be devised, or might arise from accident, of concealing the likeness between the two sounds, or, at any rate, of making it less palpable. One of such ways would be a faulty mode of spelling. If instead of ga we wrote gha the following would be the effect: the syllable would appear less simple than it really was; it would look as if it consisted of three parts instead of two, and consequently its affinity to ka would seem less than it really was. It is perfectly true that a little consideration would tell us that, as long as the sound remained the same, the relation of the two syllables remained the same; and that, if the contrary appeared to be the case, the ear was misled by the eye. Still a little consideration would be required. Now in the English language we have, amongst others, the following modes of spelling that have a tendency to mislead:—

The sounds of ph and of f, in Philip and fillip, differ to the eye, but to the ear are identical. Here a difference is simulated.

The sounds of th in thin, and of th in thine, differ to the ear, but to the eye seem the same. Here a difference is concealed.

These last sounds appear to the eye to be double or compound. This is not the case; they are simple single sounds, and not the sounds of t followed by h, as the spelling leads us to imagine.

II. Besides improper modes of spelling, there is another way of concealing the true nature of sounds. If I say that ka and ga are allied, the alliance is manifest; since I compare the actual sounds. If I say ka and gee are allied, the alliance is concealed; since I compare, not the actual sounds, but only the names of the letters that express those sounds. Now in the English language we have, amongst others, the following names of letters that have a tendency to mislead:—

The sounds fa and va are allied. The names eff and vee conceal this alliance.

The sounds sa and za are allied. The names ess and zed conceal the alliance.

In comparing sounds it is advisable to have nothing to do either with letters or names of letters. Compare the sounds themselves.

In many cases it is sufficient, in comparing consonants, to compare syllables that contain those consonants; e.g., to determine the relations of p, b, f, v, we say pa, ba, fa, va; or for those of s and z, we say sa, za. Here we compare syllables, each consonant being followed by a vowel. At times this is insufficient. We are often obliged to isolate the consonant from its vowel, and bring our organs to utter (or half utter) imperfect sounds of p', b', t', d'. In doing this we isolate the consonant.

§ 197. Let any of the vowels (for instance, the a in father) be sounded. The lips, the tongue, and the parts within the throat remain in the same position: and as long as these remain in the same position the sound is that of the vowel under consideration. Let, however, a change take place in the position of the organs of sound; let, for instance, the lips be closed, or the tongue be applied to the front part of the mouth: in that case the vowel sound is cut short. It undergoes a change. It terminates in a sound that is different, according to the state of those organs whereof the position has been changed. If, on the vowel in question, the lips be closed, there then arises an imperfect sound of b or p. If, on the other hand, the tongue be applied to the front teeth, or to the fore part of the palate, the sound is one (more or less imperfect) of t or d. This fact illustrates the difference between the vowels and the consonants. It may be verified by pronouncing the a in fate, ee in feet, oo in book, o in note, &c.

It is a further condition in the formation of a vowel sound, that the passage of the breath be uninterrupted. In the sound of the l' in lo (isolated from its vowel) the sound is as continuous as it is with the a in fate. Between, however, the consonant l and the vowel a there is this difference: with a, the passage of the breath is uninterrupted; with l, the tongue is applied to the palate, breaking or arresting the passage of the breath.

§ 198. The primary division of our articulate sounds is into vowels and consonants. The latter are again divided into liquids (l, m, n, r) and mutes (p, b, f, v, t, d, g, s, z, &c.) Definitions for the different sorts of articulate sounds have still to be laid down. In place of these, we have general assertions concerning the properties and qualities of the respective classes. Concerning the consonants as a class, we may predicate one thing concerning the liquids, and concerning the mutes, another. What the nature of these assertions is, will be seen after the explanation of certain terms.

§ 199. Sharp and flat.—Take the sounds of p, f, t, k, s; isolate them from their vowels, and pronounce them. The sound is the sound of a whisper.

Let b, v, d, g, z, be similarly treated. The sound is no whisper, but one at the natural tone of our voice.

Now p, f, t, k, s (with some others that will be brought forward anon) are sharp, whilst b, v, &c. are flat. Instead of sharp, some say hard, and instead of flat, some say soft. The Sanskrit terms sonant and surd are, in a scientific point of view, the least exceptionable. They have, however, the disadvantage of being pedantic. The tenues of the classics (as far as they go) are sharp, the mediæ flat.

Continuous and explosive.—Isolate the sounds of b, p, t, d, k, g. Pronounce them. You have no power of prolonging the sounds, or of resting upon them. They escape with the breath, and they escape at once.

It is not so with f, v, sh, zh. Here the breath is transmitted by degrees, and the sound can be drawn out and prolonged for an indefinite space of time. Now b, p, t, &c. are explosive f, v, &c. continuous.

§ 200. Concerning the vowels, we may predicate a) that they are all continuous, b) that they are all flat.

Concerning the liquids, we may predicate a) that they are all continuous, b) that they are all flat.

Concerning the mutes, we may predicate a) that one half of them is flat, and the other half sharp, and b) that some are continuous, and that others are explosive.

§ 201.—The letter h is no articulate sound, but only a breathing.

For the semivowels and the diphthongs, see the sequel.

CHAPTER II.

SYSTEM OF ARTICULATE SOUNDS.

§ 202.—The attention of the reader is now directed to the following foreign vowel sounds.

1. é fermé, of the French.—This is a sound allied to, but different from, the a in fate, and the ee in feet. It is intermediate to the two.

2. u of the French, ü of the Germans, y of the Danes.—This sound is intermediate to the ee in feet, and the oo in book.

3. o chiuso, of the Italians.—Intermediate to the o in note, and the oo in book.

For these sounds we have the following sequences: a in fate, é fermé, ee in feet, ü in übel (German), oo in book, o chiuso, o in note. And this is the true order of alliance among the vowels; a in fate, and o in note, being the extremes; the other sounds being transitional or intermediate. As the English orthography is at once singular and faulty, it exhibits the relationship but imperfectly.

§ 203. The system of the mutes.—Preliminary to the consideration of the system of the mutes, let it be observed:—

1. that the th in thin is a simple single sound, different from the th in thine, and that it may be expressed by the sign þ.

2. That the th in thine is a simple single sound, different from the th in thin, and that it may be expressed by the sign ð.

3. That the sh in shine is a simple single sound, and that it may be expressed by the sign σ (Greek σῖγμα).

4. That the z in azure, glazier (French j), is a simple single sound, and that it may be expressed by the sign ζ (Greek ζῆτα).

5. That in the Laplandic, and possibly in many other languages, there are two peculiar sounds, different from any in English, German, and French, &c., and that they may respectively be expressed by the sign κ and the sign γ (Greek κάππα and γάμμα).

With these preliminary notices we may exhibit the system of the sixteen mutes; having previously determined the meaning of two fresh terms, and bearing in mind what was said concerning the words sharp and flat, continuous and explosive.

Lene and aspirate.—From the sound of p in pat, the sound of f in fat differs in a certain degree. This difference is not owing to a difference in their sharpness or flatness. Each is sharp. Neither is it owing to a difference in their continuity or explosiveness; although, at the first glance, such might appear to be the case. F is continuous, whilst p is explosive. S, however, is continuous, and s, in respect to the difference under consideration, is classed not with f the continuous sound but with p the explosive one. I am unable to account for the difference between p and f. It exists: it is visible. It has been expressed by a term. P is called lene, f is called aspirate.

Hence p, b, t, d, k, g, s, z, are lene; f, v, þ, ð, κ, γ, σ, ζ, are aspirate. Also p, f, t, þ, k, κ, s, σ, are sharp, whilst b, v, d, ð, g, γ, z, ζ, are flat; so that there is a double series of relationship capable of being expressed as follows:—

I am not familiar enough with the early grammarians to know when the terms lene and aspirate were first used. They were the Latin equivalents to the Greek words ψίλον (psilon) and δάσυ (dasy) respectively. The Greek terms are preferable. They convey no determinate idea, whereas the Latin terms convey a false one. The origin of the word aspirate I imagine to be as follows. The Latin language, wanting both the sound of the Greek theta, and the sign to express it (θ) rendered it by th. This orthography engenders the false notion that θ differed from τ by the addition of the aspirate h. To guard against similar false notions, I rarely hereafter use the word aspirate without qualifying it by the addition of the adjective so-called.

All the so-called aspirates are continuous; and, with the exception of s and z, all the lenes are explosive.

I believe that in the fact of each mute appearing in a fourfold form (i.e. sharp, or flat, lene, or (so-called) aspirate), lies the essential character of the mutes as opposed to the liquids.

Y and w.—These sounds, respectively intermediate to γ and i (the ee in feet), and to υ and u (oo in book), form a transition from the vowels to the consonants.

§ 204. It has been seen that the sixteen mutes are reducible to four series. Of these series, p, t, k, s, may respectively be taken as the types. Of the liquids it may be predicated as follows:—

1. That m is allied to the series p.—The combination inp has a tendency to become imp.

2. That n is allied to the series t.—The combination imt has a tendency to become either impt, or int.

3. That l is allied to the series k.—The evidence of this lies deep in comparative philology.

4. That r is allied to the series s.—The evidence of this is of the same nature with that of the preceding assertion.

The series p and k have this peculiarity.—They are connected with the vowels through w and u (oo), and through y and i (ee) respectively.

§ 205. The French word roi and the English words oil, house, are specimens of a fresh class of articulations; viz., of compound vowel sounds or diphthongs. The diphthong oi is the vowel o modified, plus the semivowel y (not the vowel i) modified. The diphthongal sound in roi is the vowel o modified, plus the semivowel w (not the vowel u or oo) modified. In roi the semivowel element precedes, in oil it follows. In roi it is the semivowel allied to series p; in oil it is the semivowel allied to series k. The nature of the modification that the component parts of a diphthong undergo has yet to be determined; although it is certain there is one. If it were not so, the articulations would be double, not compound.

The words quoted indicate the nature of the diphthongal system.

1. Diphthongs with the semivowel w, a) preceding, as in the French word roi, b) following, as in the English word new.

2. Diphthongs with the semivowel y, a) preceding, as is common in the languages of the Lithuanic and Slavonic stocks, b) following, as in the word oil.

3. Triphthongs with a semivowel both preceding and following.

The diphthongs in English are four; ow as in house, ew as in new, oi as in oil, i as in bite, fight.

§ 206. Chest, jest.—Here we have compound consonantal sounds. The ch in chest is t + sh (σ), the j in jest is d + zh (ζ). I believe that in these combinations one or both the elements, viz., t and sh, d and zh, are modified; but I am unable to state the exact nature of this modification.

§ 207. Ng.—The sound of the ng in sing, king, throng, when at the end of a word, or of singer, ringing, &c. in the middle of a word, is not the natural sound of the combination n and g, each letter retaining its natural power and sound; but a simple single sound, of which the combination ng is a conventional mode of expressing.

§ 208. Other terms, chiefly relating to the vowels, have still to be explained. The é of the French has been called fermé, or close (Italian, chiuso). Its opposite, the a in fate, is open.

Compared with a in fate, and the o in note, a in father, and the aw in bawl, are broad, the vowels of note and fate being slender.

§ 209. In fat, the vowel is, according to common parlance, short; in fate, it is long. Here we have the introduction of two fresh terms. For the words long and short, I would fain substitute independent and dependent. If from the word fate I separate the final consonantal sound, the syllable, fa remains. In this syllable the a has precisely the sound that it had before. It remains unaltered. The removal of the consonant has in nowise modified its sound or power. It is not so with the vowel in the word fat. If from this I remove the consonant following, and so leave the a at the end of the syllable, instead of in the middle, I must do one of two things: I must sound it either as the a in fate, or else as the a in father. Its (so-called) short sound it cannot retain, unless it be supported by a consonant following. For this reason it is dependent. The same is the case with all the so-called short sounds, viz., the e in bed, i in fit, u in bull, o in not, u in but.

To the preceding remarks the following statements may be added.

1. That the words independent and dependent correspond with the terms perfect and imperfect of the Hebrew grammarians.

2. That the Hebrew grammars give us the truest notions respecting these particular properties of vowels.

The following sentences are copied from Lee's Hebrew Grammar, Art. 33, 34:—"By perfect vowels is meant, vowels which, being preceded by a consonant" (or without being so preceded), "will constitute a complete syllable, as בָּ‎ bā. By imperfect vowels is meant those vowels which are not generally" (never) "found to constitute syllables without either the addition of a consonant or of an accent. Such syllables, therefore, must be either like בּדַ‎ bad, or בֲּ‎ bā, i.e., followed by a consonant, or accompanied by an accent." For further remarks on this subject, see the chapter on accent.

§ 210. Before i, e, and y of the English alphabet, and before ü and ö German, the letters c and g have the tendency to assume the sound and power of s or z, of sh or zh, of ch or j; in other words, of becoming either s or some sound allied to s. Compared with a, o, and u (as in gat, got, gun), which are full, i, e, y, are small vowels.

It not every vowel that is susceptible of every modification. I (ee) and u (oo) are incapable of becoming broad. E in bed (as I have convinced myself), although both broad and slender, is incapable of becoming independent. For the u in but, and for the ö of certain foreign languages, I have no satisfactory systematic position.

§ 211. Vowel System.

From these, the semivowels w and y make a transition to the consonants v and the so-called aspirate of g (γ, not being in English), respectively.

§ 212. System of Consonants.

§ 213. Concerning the vowel system I venture no assertion. The consonantal system I conceive to have been exhibited above in its whole fulness. The number of mutes, specifically distinct, I consider to be sixteen and no more: the number of liquids, four. What then are the powers of the numerous letters in alphabets like those of Arabia and Armenia? What is the German ch, and Irish gh? Varieties of one or other of the sounds exhibited above, and not articulations specifically distinct.

§ 214. There is a difference between a connexion in phonetics and a connexion in grammar.—Phonetics is a word expressive of the subject-matter of the present chapter. The present chapter determines (amongst other things) the systematic relation of articulate sounds. The word phônæticos (φωνήτιχος) signifies appertaining to articulate sounds. It is evident that between sounds like b and v, s and z, there is a connexion in phonetics. Now in the grammar of languages there is often a change, or a permutation of letters: e.g., in the words tooth, teeth, the vowel, in price, prize, the consonant, is changed. Here there is a connexion in grammar.

That the letters most closely allied in phonetics should be most frequently interchanged in grammar, is what, on à priori grounds, we most naturally are led to expect. And that such is often the case, the study of languages tells us. That, however, it is always so, would be a hasty and an erroneous assertion. The Greek language changes p into f. Here the connexion in phonetics and the connexion in language closely coincide. The Welsh language changes p into m. Here the connexion in phonetics and the connexion in language do not closely coincide.

CHAPTER III.

OF CERTAIN COMBINATIONS OF ARTICULATE SOUNDS.

§ 215. Certain combinations of articulate sounds are incapable of being pronounced. The following rule is one that, in the forthcoming pages, will frequently be referred to. Two (or more) mutes, of different degrees of sharpness and flatness, are incapable of coming together in the same syllable. For instance, b, v, d, g, z, &c. being flat, and p, f, t, k, s, &c. being sharp, such combinations as abt, avt, apd, afd, agt, akd, atz, ads, &c., are unpronounceable. Spelt, indeed, they may be; but attempts at pronunciation end in a change of the combination. In this case either the flat letter is changed to its sharp equivalent (b to p, d to t, &c.) or vice versâ (p to b, t to d). The combinations abt, and agt, to be pronounced, must become either apt or abd, or else akt or agd.

For determining which of the two letters shall be changed, in other words, whether it shall be the first that accommodates itself to the second, or the second that accommodates itself to the first, there are no general rules. This is settled by the particular habit of the language in consideration.

The word mutes in the second sentence of this section must be dwelt on. It is only with the mutes that there is an impossibility of pronouncing the heterogeneous combinations above mentioned. The liquids and the vowels are flat; but the liquids and vowels, although flat, may be followed by a sharp consonant. If this were not the case, the combinations ap, at, alp, alt, &c. would be unpronounceable.

The semivowels, although flat, admit of being followed by a sharp consonant.

The law exhibited above may be called the law of accommodation.

Combinations like gt, kd, &c., may be called incompatible combinations.

§ 216. Unstable combinations.—That certain sounds in combination with others have a tendency to undergo changes, may be collected from the observation of our own language, as we find it spoken by those around us, or by ourselves. The ew in new is a sample of what may be called an unsteady or unstable combination. There is a natural tendency to change it either into oo (noo) or yoo (nyoo); perhaps also into yew (nyew).

§ 217. Effect of the semivowel y on certain letters when they precede it.—Taken by itself the semivowel y, followed by a vowel (ya, yee, yo, you, &c.), forms a stable combination. Not so, however, if it be preceded by a consonant, of the series t, k, or s, as tya, tyo; dya, dyo; kya, kyo; sya, syo. There then arises an unstable combination. Sya and syo we pronounce as sha and sho; tya and tyo we pronounce as cha and ja (i.e. tsh, dzh.). This we may verify from our pronunciation of words like sure, picture, verdure (shoor, pictshoor, verdzhoor), having previously remarked that the u in those words is not sounded as oo but as yoo. The effect of the semivowel y, taken with instability of the combination ew, accounts for the tendency to pronounce dew as if written jew.

§ 218. The evolution of new sounds.—To an English ear the sound of the German ch falls strange. To an English organ it is at first difficult to pronounce. The same is the case with the German vowels ö and ü and with the French sounds u, eu, &c.

To a German, however, and a Frenchman, the sound of the English th (either in thin or thine) is equally a matter of difficulty.

The reason of this lies in the fact of the respective sounds being absent in the German, French, and English languages; since sounds are easy or hard to pronounce just in proportion as we have been familiarised with them.

There is no instance of a new sound being introduced at once into a language. Where they originate at all, they are evolved, not imported.

§ 219. Evolution of sounds.—Let there be a language where there is no such a sound as that of z, but where there is the sound of s. The sound of z may be evolved under (amongst others) the following conditions. 1. Let there be a number of words ending in the flat mutes; as slab, stag, stud, &c. 2. Let a certain form (the plural number or the genitive case) be formed by the addition of is or es; as slabis, stages, studes, &c. 3. Let the tendency that words have to contract eject the intermediate vowel, e or i, so that the s of the inflexion (a sharp mute) and the b, d, g, &c. of the original word (flat mutes) be brought into juxta-position, slabs, studs, stags. There is then an incompatible termination, and one of two changes must take place; either b, d, or g must become p, t, or k (slaps, staks, stuts); or s must become z (stagz, studz, slabz). In this latter case z is evolved. Again,

Let there be a language wherein there are no such sounds as sh, ch (tsh), or j (dzh); but where there are the sounds of s, t, d, and y.

Let a change affect the unstable combinations sy, ty, dy. From this will arise the evolved sounds of sh, ch, and j.

The phenomena of evolution help to determine the pronunciation of dead languages.

§ 220. On the value of a sufficient system of sounds.—In certain imaginable cases, a language may be materially affected by the paucity of its elementary articulate sounds.

In a given language let there be the absence of the sound z, the other conditions being those noted in the case of the words stag, slab, stud, &c. Let the intermediate vowel be ejected. Then, instead of the s being changed into an evolved z, let the other alternative take place; so that the words become staks, slaps, stuts. In this latter case we have an alteration of the original word, brought about by the insufficiency of the system of articulate sounds.

§ 221. Double consonants rare.—It cannot be too clearly understood that in words like pitted, stabbing, massy, &c. there is no real reduplication of the sounds of t, b, and s, respectively. Between the words pitted (as with the small-pox) and pitied (as being an object of pity) there is a difference in spelling only. In speech the words are identical. The reduplication of the consonant is in English, and the generality of languages, a conventional mode of expressing upon paper the shortness (dependence) of the vowel that precedes.