In the first place, then, our automatisms are _independent_ phenomena; they are what the physician calls _idiognomonic_. That is to say, they are not merely symptomatic of some other affection, or incidental to some profounder change. The mere fact, for instance, that a man writes messages which he does not consciously originate will not, when taken alone, prove anything beyond this fact itself as to the writer's condition. He may be perfectly sane, in normal health, and with nothing unusual observable about him. This characteristic--provable by actual observation and experiment--distinguishes our automatisms from various seemingly kindred phenomena. Thus we may have to include in our class the occasional automatic utterance of words or sentences. But the continuous exhausting vociferation of acute mania does not fall within our province; for those shouts are merely _symptomatic_; nor, again, does the _cri hydrocéphalique_ (or spontaneous meaningless noise which sometimes accompanies water on the brain); for that, too, is no independent phenomenon, but the direct consequence of a definite lesion. Furthermore, we shall have to include in our class certain simple movements of the hands, co-ordinated into the act of writing. But here, also, our definition will lead us to exclude _choreic_ movements, which are merely symptomatic of nervous malnutrition; or which we may, if we choose, call _idiopathic_, as constituting an independent malady. But our automatisms are not _idiopathic_ but _idiognomonic_; they may indeed be associated with or facilitated by certain states of the organism, but they are neither a symptom of any other malady, nor are they a malady in themselves. Agreeing, then, that our peculiar class consists of automatisms which are idiognomonic,--whose existence does not necessarily imply the existence of some profounder affection already known as producing them,--we have still to look for some more positive bond of connection between them, some quality common to all of them, and which makes them worth our prolonged investigation.