¹ Emsley has noted Types I and III, but he omits Type II, since it was not important for his purpose.

² A survey of English dictionaries; London, 1933.

³ Am. Sp. 15.57 (1940).

⁴ Mathews states that Bailey gave aid, ‘though of an elementary sort, in syllabification and pronunciation’ (Survey of English dictionaries 30). The amount of aid given to pronunciation has been discussed above. It is evident that Bailey's system gives no help in syllabification, since he does not mark syllables. Halcyon, for instance, pronounced in two or in three syllables during the 18th century, is marked ha'lcyon in Bailey 1730; does this call for two syllables or for three?

⁵ Mathews 28.

⁶ Not all later editions of Bailey adopt this system. The edition of 1755, entitled A New Universal Etymological English Dictionary, revised and corrected by J. N. Scott, keeps the old accentual system of the 1730 and 1736 editions of Dictionarium Britannicum. On the other hand, Buchanan's revision of 1760, A New Etymological English Dictionary, employs the revised Bailey system. This was three years after the appearance of Buchanan's own dictionary, containing the first truly diacritical notation.

⁷ This word is interesting for another reason. The pronunciation of the a was controversial in the 18th century, pronounced like the a of hat or the a of fate. Bailey's notation probably indicates short a (brass-), but if it were taken literally, by a foreigner let us say, it really indicates a long pronunciation of the silent e (-see 'let). This word illustrates well the inaccuracy of Type II.

⁸ He apparently does not mind using the accent mark at ends of words after vowels, for he has rely', reply'.

⁹ Buchanan's pronunciations of initial unstressed a seem overprecise, as well as incomprehensibly varied.

¹⁰ See particularly the varieties of symbols used in Webster, in Winston, and in the Standard for the unaccented neutral vowel of sofa and other words.

¹¹ In examining Perry, I have had to use a 1788 American edition from the 4th British edition. This, however, contains the 1777 preface with Perry's discussion of his vowel system. The entries correspond to the system and principles laid down for vowels in this preface. But in the preface there is no description of consonant markings. I have no way of knowing whether changes in these had been introduced between 1777 and 1788.

¹² Thus bravado is entered bra-vā'do (no accent over first a or o, indicating therefore a shorter duration of long a and o).

¹³ What is most interesting here is that Perry in his 1777 Preface (vi) criticizes adversely Sheridan's proposed key for containing ‘several duplicates of the same sound, differently marked.‘ Perry's excuse for falling into this same error might have been that, whereas Sheridan proposed to respell, he did not.

¹⁴ See footnote 12.

¹⁵ It need not be stressed here how unreliable as phoneticians these early dictionary makers were. Contemporaries of Perry—Sheridan and Walker—both use different symbols for these two vowels, although both remark in their introductions that one is simply the short sound of the other.

¹⁶ It is true, however, that Kenrick uses it very rarely, probably too rarely to give an honest picture of contemporary usage.

¹⁷ This is no academic question to the student, for both pronunciations of conduit were starting the struggle for supremacy at this very time.

¹⁸ Of Sheridan, Walker says: ‘It must, indeed, be confessed, that Mr. Sheridan's Dictionary is greatly superiour to every other that preceded it; and his method of conveying the sound of words, by spelling them as they are pronounced, is highly rational and useful.—But here sincerity obliges me to stop. The numerous instances I have given of impropriety, inconsistency, and want of acquaintance with the analogies of the Language, sufficiently show how imperfect I think his Dictionary is’ (Preface iii). In speaking of impropriety, etc., Walker is referring to the pronunciations recommended, not to the pronouncing system.

¹⁹ It is interesting to compare Kenrick, Perry, and Scott here with Sheridan for consonant representation: Kenrick ac-cess, Perry ac-cess, Scott ac cess.

²⁰ The e of bet is used in -er- words: ceremony, guerdon, mercenary, verjuice, etc.

²¹ Walker adds that in his opinion yello is ‘more agreeable to analogy, and the best usage’.