altogether of words which had a definite meaning attaching to them even when uttered by themselves, and consequently that the words of the second class grew, so to say, out of the words of the first class. These are the conclusions which a mere examination of a single language, our own, under the guidance of observation and common sense, would force upon us; always supposing our language to be a representative one. And these conclusions are strengthened when we come to look a little into the history of words, so far as we can trace it. So far back, therefore, we may go in the history of language to a time when all the words which men used were words which by themselves evoked distinct ideas. Relegating these words, as far as we can, into the classes which grammarians have invented for the different parts of speech, we see that the significant words are all, as a rule, either nouns (or _pro_-nouns), adjectives, or verbs; that the insignificant words are, as a rule, adverbs, prepositions, and conjunctions--what, in fact, are called _particles_, fragments of speech. I say, _as a rule_, for both divisions. The pronouns and the auxiliary verbs, for example, are very difficult to classify; and it depends rather on their use in each individual sentence, to which division they are to be relegated. [Sidenote: Origin of speech undiscoverable.] But though we have now learnt to distinguish the words which by themselves convey definite ideas, and those others whose meaning depends upon the first class, we are as far as ever from understanding how words, whether of one kind or the other, come to have the significance which they have for us. _Book_--no sooner have we pronounced the word than an _idea_ more or less distinct comes into our mind. The thought and the sound seem inseparable, and we cannot remember the time when they were not so. Yet the connection between the thought and the sound is not necessary. In fact, a sound which generally comes connected with one idea may--if we are engaged at the time upon a language not our own--enter our minds, bringing with it an idea quite unconnected with the first. _Share_ and _chère_, _plea_ and _plie_, _feel_ and _viel_ (German), are examples in point; and the same thing is shown by the numerous sounds in our language which have two or more quite distinct meanings, as for example--_ware_ and _were_, and (with most people) _where_ too. _Rite_ and _right_ and _wright_ are pronounced precisely alike; therefore there can be no reason why one sound should convey one idea more than another. In other words, the idea and the sound have an arbitrary, not a natural connection. We have been _taught_ to make the sound ‘book’ for the idea book, but had we been brought up by French parents the sound ‘livre’ would have seemed the natural one to make.