204 The Loom of Language battery of new agglutinative contractions for the definite article (p 361) accompanied by a preposition Like other formative processes, levelling or regulan/ation by analogy waxes m periods of illiteracy and culture contact, waning under the discipline of script The part it has played m the evolution of our remaining flexions will come up for further discussion m Chapter VI What applies to flexions, or to derivative affixes such as the -cr in baker^ applies equally to pronunciation, to word order and to syntax m general Habit, local or personal limitations of vocabulary and human la/iness continually conspire to impose the pattern of the more familiar word or phrase on those we use less often To the extent that grammarians have set themselves against the populai drift towards (pp 168 and 267) regularity, their influence has been retrograde Analogical extension is the process by which natural languages are always striving to assume the orderliness of a constructed auxiliary To get rid of the disorder inherent m natural languages was the cardinal motif of language planning m the lattei half oi the nineteenth century The issue was not entirely novel The grammarians of antiquity had discussed it and were of two minds One party, the wioniah\U9 took the conservative view The other, the analogies swam with the stream, and even practised revision of texts to prune away grammatical irregu- lanties The contioversy went on for several centuries Among others.* Julius Caesar took a hand in it, As a general he favoured regimentation So he naturally took the side of the analogies, The fact that isolation is tJhe predominant feature of some languages (e g. Chinese dialects or Malay), regularity of affixes the outstanding characteristics of others (c,g, Finno-Ugnan dialects, Japanese, Turkish) and chaotic irregularity of suffixes the prevailing grammatical pattern of a thiid group (e g Sanskrit, Greek, Laim or Old English) has prompted speculations which take us into the twilight of human speech, without much hope of reaching certainty. Some linguists believe that primitive speech was a sing-song matrix from which words emerged with the frayed edges of a Sanskrit noun or verb, According to this view there has been a steady progress from amalgamation, through agglutinative regularity to isolation. Others favour the opposite view. They believe that the speech of our primitive ancestors once consisted of separate root^words which were probably monosyllabic, like those of Chinese dialects* If so* words which carried less emphasis than others beqaxae attached as modifiers to more meaningful ones. Finally, these accretions got intimately fused, and forfeited their former independence.