And these sounds may be measured and classified to a certain extent according to the emotions which cause them, although it must be borne in mind that we are looking at the matter collectively; that is to say, without reckoning on individual idiosyncrasies of expression in speech. Of course we know that joy is apt to make us raise the voice and sadness to lower it. For instance, we have all heard gruesome stories, and have noticed how naturally the voice sinks in the telling. A ghost story told with an upward inflection might easily become humourous, so instinctively do we associate the upward inflection with a non-pessimistic trend of thought. Under stress of emotion we emphasize words strongly, and with this emphasis we almost invariably raise the voice a fifth or depress it a fifth; with yet stronger emotion the interval of change will be an octave. We raise the voice almost to a scream or drop it to a whisper. Strangely enough these primitive notes of music correspond to the first two of those harmonics which are part and parcel of every musical sound. Generally speaking, we may say that the ascending inflection carries something of joy or hope with it, while the downward inflection has something of the sinister and fearful. To be sure, we raise our voices in anger and in pain, but even then the inflection is almost always downward; in other words, we pitch our voices higher and let them fall slightly. For instance, if we heard a person cry "Ah/" we might doubt its being a cry of pain, but if it were "Ah\" we should at once know that it was caused by pain, either mental or physical. The declamation at the end of Schubert's "Erlking" would have been absolutely false if the penultimate note had ascended to the tonic instead of descending a fifth. "The child lay dead."