New necessities, new evolutions of society into more complex conditions, evoke new words; which come forth, because they are required now; but did not formerly exist, because they were not required in the period preceding. For example, in Greece so long as the poet sang his own verses ‘singer’ (ἀοιδὸς) sufficiently expressed the double function; such a ‘singer’ was Homer, and such Homer describes Demodocus, the bard of the Phæacians; that double function, in fact, not being in his time contemplated as double, but each part of it so naturally completing the other, that no second word was required. When, however, in the division of labour one made the verses which another chaunted, then ‘poet’ or ‘maker’, a word unknown in the Homeric age, arose. In like manner, when ‘physicians’ were the only natural philosophers, the word covered this meaning as well as that other which it still retains; but when the investigation of nature and natural causes detached itself from the art of healing, became an independent study of itself, the name ‘physician’ remained to that which was as the stock and stem of the art, while the new offshoot sought out a new name for itself. Another motive to the invention of new words, is the desire thereby to cut short lengthy{122} explanations, tedious circuits of language. Science is often an immense gainer by words, which say singly what it would have taken whole sentences otherwise to have said. Thus ‘isothermal’ is quite of modern invention; but what a long story it would be to tell the meaning of ‘_isothermal_ lines’, all which is summed up in and saved by the word. We have long had the word ‘assimilation’ in our dictionaries; ‘dissimilation’ has not yet found its way into them, but it speedily will. It will appear first, if it has not already appeared, in our books on language{123}. I express myself with this confidence, because the advance of philological enquiry has rendered it almost a matter of necessity that we should possess a word to designate a certain process, and no other word would designate it at all so well. There is a process of ‘assimilation’ going on very extensively in language; it occurs where the organs of speech find themselves helped by changing a letter for another which has just occurred, or will just occur in a word; thus we say not ‘_adf_iance’ but ‘_aff_iance’, not ‘re_n_ow_m_’, as our ancestors did when the word ‘renommée’ was first naturalized, but ‘re_n_ow_n_’. At the same time there is another opposite process, where some letter would recur too often for euphony or comfort in speaking, if the strict form of the word were too closely held fast, and where consequently this letter is exchanged for some other, generally for some nearly allied; thus it is at least a reasonable suggestion, that ‘cœ_r_uleum’ was once ‘cœ_l_uleum’, from cœlum: so too the Italians prefer ‘ve_l_e_n_o’ to ‘ve_n_e_n_o’; and we ‘cinnamo_n_’ to ‘cinnamo_m_’ (the earlier form); in ‘turtle’ and ‘purple’ we have