50 The Loom of Language They came into use about three thousand years ago, but the inherent possibilities of an invention which we now recognize as one of the out- standing cultural achievements of mankind incubated slowly during the course of successive millcma The first peoples who used alpha- betic writing did so for short inscriptions in which individual letters might be written upside down or reversed sideways,, with little con- sideration for the reader (Fig 38) Even when a secular literature spread through the Greek and Roman world., the written language remained a highly artificial product remote from daily speech, Greek writing was never adapted to rapid reading.) because Greek scribes never consistently separated words. The practice of doing so did not become universal among Roman writers. It became a general custom about the tenth century of oui own era When printing began3 crafts- men took pride in the ready recognition of the written word, and punctuation marks., which individual writers had used sporadically without agreement, came into their own Typographers first adopted an agreed system of punctuation, attributed to Aldus Manutius^ in the sixteenth century In the ancient world the reader had to be his own palaeographer To appreciate the gap between modem and ancient reading, compare the sentences printed below. KINGCHAiaESWALiODANDTAM King Charles walked and talked Half an hour after his head was cut oil To do justice to the story of the alphabet we must start by examining the meaning of a few technical terms. Word is itself a technical term It is not easy to define what we mean by a separate word in all cir- cumstances So let us imagine what a tiaveller would do if he came to live with an illiterate tribe in the interior of Borneo By pointing at things around he might soon learn which sounds stand for fiidurablc objects. By comparing similar things he might also learn to recogni7e sounds signifying qualities such as red> rough> or round By watching people together he could also detect sounds which are signals of action like James! Here! Cornel Hurry! All this would not make a complete inventory of the elements of a continuous conversation If the language contained words corresponding to and9 during mean- while, for^ or according, he would take a long while to decide how to use them., because they never stand by themselves For the same reason it would also be difficult to decade whether to regard them a$ separate words.