All consonants are pronounced with a less cavity than any of the vowels; and consequently they contribute to form a sound still more sharp than the sharpest vowel pronounced single. Hence it follows, that every articulate sound into which a consonant enters, must necessarily be double, though pronounced with one expiration of air, or with one breath as commonly expressed. The reason is, that though two sounds readily unite; yet where they differ in tone, both of them must be heard if neither of them be suppressed. For the same reason, every syllable must be composed of as many sounds as there are letters, supposing every letter to be distinctly pronounced. We next inquire, how far articulate sounds into which consonants enter, are agreeable to the ear. With respect to this point, there is a noted observation, that all sounds of difficult pronunciation are to the ear harsh in proportion. Few tongues are so polished as entirely to have rejected sounds that are pronounced with difficulty; and such sounds must in some measure be disagreeable. But with respect to agreeable sounds, it appears, that a double sound is always more agreeable than a single sound. Every one who has an ear must be sensible, that the diphthongs _oi_ or _ai_ are more agreeable than any of these vowels pronounced singly. And the same holds where a consonant enters into the double sound. The syllable _le_ has a more agreeable sound than the vowel _e_ or than any vowel. And in support of experience, a satisfactory argument may be drawn from the wisdom of Providence. Speech is bestowed upon man, to qualify him for society. The provision he hath of articulate sounds, is proportioned to the use he hath for them. But if sounds that are agreeable singly were not also agreeable in conjunction, the necessity of a painful selection would render language intricate and difficult to be attained in any perfection. And this selection, at the same time, would tend to abridge the number of useful sounds, so as perhaps not to leave sufficient for answering the different ends of language. In this view, the harmony of pronunciation differs widely from that of music properly so called. In the latter are discovered many sounds singly agreeable, that in conjunction are extremely disagreeable; none but what are called _concordant sounds_ having a good effect in conjunction. In the former, all sounds singly agreeable are in conjunction concordant; and ought to be, in order to fulfil the purposes of language.