Accidence—The Table Manners of Language 93 Thus an ordinary English dictionary which contains behave and beha- viour > does not list behaved., behaves^ or behaving The part of grammar called accidence consists of rules for detecting how to form such deriva- tives and how they affect the meaning of a dictionary word which shares the same root Our first task must therefore be to recall (p. 53) how single words can grow First of all, they can do so by fusing with one another or with meaning" ful affixes (a) Because the meaning of the compound word (e g brickyard) so formed is sufficiently suggested by the ordinary meaning of its separate parts in a given context This is a trick specially charac- teristic of Teutonic languages, Greek and Chinese (b) Because two native words constantly occur in the same context and get glued together through slipshod pronunciation, as in the shortened forms dont, wont, cant, shant for do not, will not, can not, diall not, as also don (= do on) and doff (= do off) (c) Because an affix (p 53) borrowed from another language is attached to them, as the Latin ante- (before) is used m antenatal chnic, or the Greek anti- (against) in anti-fascist, anti-comintern, and <272r/-anything-else-which-we-do-not-hke It is useful to distinguish fusion due to speech-habits, i e. (b) from fusion associated with meaning, i.e. (a) and (c). The word agglutination refers to the former, i.e. to fusion arising from context and pronunciation without regard to meaning Once fusion has begun another process begins to woik. The meaning like the form of a word part becomes blurred People get careless about the meaning of an affix. We expect a word to end (or to begin) in the same way, when we have made a habit of using similar words with the same affix in a similar context. This leads to a habit of tacking on the same affix to new words without regard to its original meaning Having made a word mastodon, we add the -s of mastodons because we are used to treating animals in this way What grammarians call analogical extension includes this process of extending the use of an affix by analogy with pre-existing words built up m the same way Children and immigrants (see p. 168), as well as native adults, take a hand in the way languages change for better or for worse. For instance, an American or British child who is accustomed to saying / caught^ when he means that he has made his catch3 may also say the eggs haught for the eggs hatched^ or, being more accustomed to adding ~ed9 may say / catched for / caught. This process is immensely important (see p 203) in building up new words or in changing old ones. We should therefore recognize its limitations at the outset.