BOAS] HANDBOOK OF AMREICAN INDIAN LANGUAGES 21 



readily differentiated. The strongest of these we call the "fortis," 

 and indicate it by following the consonant by an ! {jp!, t!). 



When these stops are not accompanied by onj kind of vibra- 

 tion of the vocal chords, they are called "surds." 



It is, of course, also possible that more than one stop may be made 

 at one time. Thus it might be possible to close at the same time 

 the hps and the posterior part of the mouth with the tongue. Tliis 

 type of combination is, however, rare; but we find very frequently 

 articulation of the vocal chords with stops. This results in the 

 voiced consonants, or sonants. In English we find that almost 

 always the stress of articulation of the voiced sound is less than the 

 stress of articulation of the unvoiced sound, or surd; but tliis cor- 

 relation is not necessary. In American languages particularly, we 

 find very commonly the same degree of stress used with voicing 

 and without voicing, which brings it about that to the European ear 

 the surd and sonant are difficult to distinguish. 



A tliird modification of the consonants is brought about by the 

 strength of breathing accompanying the release of the closure. In a 

 sound like t, for instance, the sound may be simpl}^ produced by 

 closing the mouth, by laying the tip of the tongue firml}^ against 

 the palate, producing a slightly increased amount of air-pressure 

 beliind the tongue, and then releasing the closure. On the other 

 hand, the sound may be produced by bringing about the closure 

 and combining the release with the expiration of a full breath. 

 Sounds which are accompanied by tliis full breathing may be called 

 "aspirates," and we will designate the aspiration by ', the symbol 

 of the Greek spiritus asper. Tliis full breathing may follow the 

 stop, or may begin even before the completion of the closure. With 

 the increased stress of closure of the fortis is connected a closure of 

 the glottis or of the posterior part of the tongue, so that only the air 

 that has been poured into the vocal cavity is expelled. 



In the case of voiced consonants, the voicing may either be en- 

 tirely synchronous with the consonant, or it may slightly precede or 

 follow it. In both of these cases we may get the impression of a 

 preceding or following exceedingly weak vowel, the timbre of wliich 

 will depend essentially upon the accompanying consonant. When 

 the timbre is very indefinite, we ^vrite tliis vowel E; when it is more 

 definite, A, I, 0, U, etc. In other cases, where the release at the 



