tendency-which may be observed in English also-to make the pronunciation not only of oratory but of mere reading aloud distinct from that of everyday life, as is shown very clearly in the liaisons. Thus, in reading aloud, a Frenchman would sound the (t) of the ending -ment before a vowel, but never in speaking.

Here the principle of association comes in. To a Frenchman the ending -ment suggests primarily the pronunciation (-mã) before a vowel as well as a consonant; but when he speaks or reads to an audience, he makes an effort to sound the (t) before a vowel; just as an Englishman in speaking slowly and solemnly may make (kaant) into (kæn not), although in English there is no necessity felt for departing from the colloquial pronunciation. It is evident that the first and most immediate associations of the foreign learner ought to be with the colloquial forms. When he has learnt these, he will be on a level with the educated native, and, like him, can afterwards learn the more artificial pronunciation, and thus establish a series of secondary oratorical associations. If his associations are primarily with the oratorical forms, his ordinary conversation will be unnatural and offensive to the native ear.

Degrees of Colloquialism

But there are degrees of colloquialism. In all languages the pronunciation may vary according to the degree of familiarity between the speakers. Even in England a young man will sometimes unconsciously modify his pronunciation in speaking to a strange lady or an older man.

The mood of the speaker, too, may have an effect. Tension of mind-as in giving definite directions, explaining a difficulty, impatient command-is naturally accompanied by greater vigour of enunciation; while indifference and languor show themselves in half-finished consonants and curtailed soundgroups. We can hear in English the sharp snap of what! degenerate in the mouth of the same speaker into the languid (woh) or almost (waa), which may further degenerate into a mere grunt.

same person may vary This is very marked in

Again, the pronunciation of the according to the speed of utterance. French, where the elimination of the weak (ə) depends greatly In Passy's Elementarbuch the texts are given in the

on speed.

pronunciation of medium speed, a quicker and a slower pronunciation being occasionally given in the notes. Thus to the normal (õ vjẽ d sɔne msjø) and (i j ãn a dǝ tut le kulœœer) correspond the slow (õ vjẽ də sɔne mǝsjø) and the quick (j ăn a d tut le kulcecer), and to the medium (estrǝrdinɛɛr) (si vu plɛ), the slow (ekstraordinɛɛr) and the quick (sj u plɛ).

It is evident that the foreigner should aim at what may be called a medium colloquial style of pronunciation. It is painful and incongruous to hear the rapid pronunciation of clipped speech reproduced in a slow, solemn, oratorical tempo. On the other hand, it is much more irrational to teach a foreigner pronunciations which never occur in the colloquial speech of natives. The best general advice is therefore: never be oratorical; be colloquial, but not too colloquial.

The revolt against artificial standards of pronunciation sometimes tempts phonetic enthusiasts into constructing colloquial monstrosities when dealing with a foreign language-they become more colloquial than the most slovenly native. Thus a foreigner who has learnt to obscure weak-stressed vowels in English-who has learnt to say (kæriktə, maagit, izri-əl) in spite of the associations of the written forms character, Margate, Israel-is apt to get reckless, and go too far in this direction, making perhaps (nɔ'wijdzǝn næpsæk) into (nǝwijdzǝn næpsək), pronunciations which I remember having seen actually given.

Vulgarisms should be avoided; not because they are in themselves ugly or less logical, or in any way more objectionable than the corresponding polite forms, but simply because they belong to a different dialect. But we must distinguish between real and theoretical vulgarisms: that is, between forms which, as a matter of fact, do not occur in educated speech, and those which are commonly called 'vulgar,' and yet do occur in educated speech. Of theoretical vulgarisms, some are simply universal in educated speech, such as the loss of the consonant (r) in lord by which this word becomes identical in pronunciation with laud, others widely spread, such as the (r) in idea(r) of, India(r) Office. But as this latter colloquialism is not universal, the insertion of the (r) generally occurring only in rapid speech and in closely connected groups of words, so that its omission does not produce any effect of unreality or artificiality, it would be mere perversity in the foreigner to imitate it in his slow pronunciation. But while it is a real vulgarism to omit (h) in full-stressed words, it is a disagreeable affectation

1

not to drop it in such collocations as tell him. This affectation is widely spread; but it is always artificial; so that the speakers who try to keep it up consistently are always liable to fail. For these reasons a foreigner should avoid it: that is, he should say (tel -im), keeping the (h) for the emphatic (tel him not həə).

The statements of unphonetic natives about vulgarisms and other varieties of pronunciation are never reliable, and should be listened to with great caution. A foreigner once asked a learned Englishman which was right, (aast) or (aaskt), as the preterite of ask; and was told that there was no such pronunciation as (aast). A minute after the learned man was heard to say (sou ij aast im ən aast im ən aast im ǝgen). On another occasion a well-known authority on the English language began in a mixed company to denounce the vulgarisms in my Elementarbuch des gesprochenen Englisch. A German pupil of mine who was present sent a whisper round the circle, telling them to listen carefully for these very vulgarisms in the authority's own pronunciation. The latter then began a lengthy harangue; and, to his surprise, was continually interrupted by bursts of laughter from his audience.

Standards of Pronunciation

As the educated pronunciation of a language is never absolutely uniform, the question arises, which is the standard? To the foreigner this is not a sentimental or æsthetic question, but a purely practical one.

As the literary languages of most countries are simply the fossilized dialects of their respective capitals-literary French being nothing but the written form of the older Parisian dialect, literary English of the older London dialect-there seems every reason why the dialect of the capital should be taken as the standard of the spoken language as well. Practical considerations point to the same view. First, there is the numerical preponderance of the speakers of the dialect of the capital. Secondly, foreigners naturally gravitate to the capital, or, at any rate, make it their starting-point. Even in Germany, where there is much less centralization than in France and England, it is surely more practical for the foreigner to learn the educated speech of Berlin than that of some provincial town where on abstract grounds 'the best German' is said to be spoken.

Even within the narrowest limits there may be differences of pronunciation. Even in educated Southern English we sometimes find a word pronounced in several ways. When Dr. Johnson was asked by a lady whether he pronounced the word neither as (naiðǝr) or (niiðǝr), he replied (needǝr, mædæm). The last pronunciation is now extinct, but the other two still seem to be about equally frequent. The fluctuations of French pronunciation are even greater. In such cases the learner must select one pronunciation and keep to it. It follows, of course, that his text-books should, as far as possible, give a uniform pronunciation, no matter how arbitrary the selection may be.

Pronunciation of Rare Words

For rare words which the learner meets for the first time in nomic texts, he will require a pronouncing dictionary. Such a dictionary may be shortened and made more convenient by the omission of all the commoner words which the learner who has read a few phonetic texts cannot help knowing thoroughly.

The learner should not be too scrupulous about the pronunciation of rare foreign words in the language he is studying, such as barbarous geographical names, which may fill the newspapers for a few weeks, and then be quite forgotten. When a foreigner wants to know exactly how such a name as Ujiji ought to be pronounced, he should be told to guess at it by analogy, taking care not to anglicize it-in fact, to do what an English reader would do with an unfamiliar word he had never heard spoken, but only seen in print. When a foreigner reflects that such a word as Zulu is not pronounced (zjuwljuw) but (zuwluw), he must see that it would be contrary to analogy to give the first i in Ujiji the English value (ai); it must be either (ij) or (i)—it does not matter which. Such a word cannot have a fixed traditional pronunciation.

In introducing words from our own language into the foreign language we are speaking, we must be careful about trying to adapt its pronunciation to that of the foreign language; where there is doubt, it is safest to keep the native pronunciation unchanged. I remember having constantly to correct a Norwegian who pronounced the name of the Norwegian town Bergen as (bəǝdzən). I told him that if he must anglicize it, let him call it (bɛəgən), which would be the average educated

L

Englishman's imitation of the native pronunciation. So also, when an Englishman uses such a German name as Beethoven in speaking French, it is much safer to keep the German pronunciation than to try and make up a French pronunciation with a final nasal vowel.

If, on the other hand, a native name has two pronunciations, one of which agrees with the spelling, the latter is generally sure to be the most modern one, and should therefore be adopted by a foreigner, who, for instance, will find himself on the side of the increasing majority if he pronounces such names as Cirencester and Abergavenny as they are written. If he does the same with Coke, Home, Cowper, instead of calling them (kuk, hjuwm, kuwpǝ), he will at least have many mispronouncers on his side.