Again, the first argument in favour of Waitz’s theory is itself clearly founded upon a paradox. It can scarcely be seriously maintained that while we can trace the growth of implements, such as spears and knives, from the simplest possible form upwards, such implements as speech and social laws have been ready made in a highly complex form. Argument number two serves to expose the grossness of this paradox. It would be as reasonable to maintain that mankind had begun by drawing pictures before they learnt to draw the elements out of which the pictures were composed. The whole theory, therefore, belongs to the category of theories which explain _obscurum per obscurius_. It may be, and no doubt is, practically impossible to explain in any _natural_ way how speech arose. But at all events it is easier to understand how it may have arisen in a simple form and grown to one more complex, than to imagine it beginning in a complex state and by detrition resolving into simple elements. P. 68. _Consonantal and vowel sounds._--The fact that even in Aryan roots the consonants have more weight than the vowel sounds will be evident merely from the instances given in the course of this and the following chapter--_fly_, _flee_, _flew_ (_w_ is here a vowel sound); _night_, _Nacht_; _knight_, _Knecht_; _Raum_, _room_; _asmi_, _esmi_ (_eimi_), _sum_, etc. This general rule holds good for almost all languages, and seems necessarily to do so from the stronger character of the consonantal and the weaker character of the vowel sounds. But the _relative_ importance of vowels and consonants is very different in different classes of language. In the Aryan tongues the essential root is made up of vowels and consonants, and the variations upon the root idea are _generally_ expressed by additions to the root and not by internal changes in it. In this way, as we saw, most grammatical inflexions are made: hom-o, hom-inis, am-o, am-abam, τύπτω, ἒτυτον, ἒτυψον, etc. But in Semitic languages the root consists of the consonants only, and the inflexions are produced by internal changes, changes of the vowels which belong to a consonant. For example, in Arabic the three consonants _k-t-l_ (_katl_) represent the abstract notion of the act of killing. From them we get _kátil_, one who kills; _kitl_ (pl. _aktal_), an enemy; _katala_, he slew; _kutila_, he was slain. From _z-r-b_ (_zarb_), the act of striking; _zarbun_, a striking (in concrete sense); _zarábun_, a striker; _zaraba_, he struck; _zuriba_, he was struck. Compare these with occido, occidi, occisor, or with τύπτω, τέτυφα, etc., and we see that in the Aryan tongues the radical remains almost unchanged, and the inflexions are made _ab extra_; but in the Semitic language the inflexions are made by changes of vowel sound within the framework of the root consonants.