Comparison - A Dawning Light 1.7 2006/04/21 22:35:56 GMT-5 2006/05/27 02:12:20.318 GMT-5 Joseph E. Grimes joe_grimes@sil.org Joseph E. Grimes joe_grimes@sil.org comparative linguistics correspondences historical linguistics word lists Wordcorr Sherlock Holmes examines the word lists, notices patterns in them, and attempts analysis using a model of Charles Babbage's Analytical Engine. Second of three Comparison modules. When I came down to breakfast the next morning, Holmes was at his desk, a pile of foolscap in front of him and other sheets lying crumpled on the floor. “Well, Holmes,” I said. “Aren't you stirring rather early?” “Rather late, Watson. I started looking at the lists of words Mr Bond left us. Though the languages show a great similarity, deuce take me if I can find any consistency to it! I've been at it all night.” “Then I prescribe a couple of hours' quick sleep, followed by a hearty breakfast of eggs, ham, and coffee. I know you won't depart from the chase for any longer than that, but it will clear your brain. I'll ask Mrs Hudson to lay breakfast for you at half past nine.” He made himself comfortable on the couch. When he slept, I made arrangements for his breakfast. Then I took up Bond's lists of words, and soon saw for myself that for whatever pattern I thought I discerned, an exception to it lay close at hand. Holmes awoke in due time. While we breakfasted, he laid out his frustrations to me. “Beastly things, Watson, these foreign languages! Yet I suppose were our own English laid alongside Dutch and German, we should find the same tangle. There is a consistency here, I swear; but it's woven through so many threads that even the most careful note-taking fails to capture it. If only I could follow thirty or forty trails at once! “For example, look at the first entry on the list. In all three languages the words begin with a p sound. From that we can deduce nothing about how the languages diverged, though it does suggest that the progenitor language used the p sound. “Examples like ka'daro 'husk' and bunga 'flower', in which all the sounds match perfectly in all three languages, also tell us nothing about differences, but they do expand the list of sounds that occur at the beginning, middle, or end of words, especially if we find the same correspondences elsewhere in the list. “Some of the entries don't seem to point to a common ancestor at all. 'Leaf,' for example, is leko' in Makassar, daung in Bugis, and taha in Saleyer. Since there are no similarities, it could indicate that the people Bond interviewed were simply thinking of different kinds of leaves. They might well have in their vocabularies words that match those of the other languages, but those words didn't come to mind in the situation. Or some of the words might be taken from still other languages, much as we English use words taken from French every day, to the disgust of the French. “But the entries where one word is similar to the others, yet not identical to them, raise interesting questions of how the languages may have changed with time. Look at the entry for 'banana'; the sounds nt in Makassar unti match a lengthened tt in Bugis utti, where the n sound in that position may have taken on the properties of its neighbor. We don't know where Saleyer stands on that one, because they use a different word, but as we comb through more entries, that may become clear. “The entry for 'grass' suggests the same sort of thing: Rea in Makassar and Saleyer, area in Bugis. Either Bugis has added a prefix a-, or an initial a has disappeared, so we must keep on the lookout for other evidence of the same process.” “But Holmes,” I exclaimed, “you've said several times already that we need to track similar kinds of change through many cases, and every one of them may seem equally tenuous!” “Tenuous, yes, my friend. But though one correspondence by itself be light as gossamer, enough similar cases may weave a strong fabric indeed,” said Holmes. I answered, “Then that must be what kept you up all night, the thirty or forty trails you spoke of trying to follow.” “Exactly, Watson. And now that you mention it, a new thought comes to mind! “What if we could harness the power of Babbage's analytical engine to keep track of all that for us! Quick, write off some notes for yourself on what we have been saying, while I prepare to make off with the whole manuscript to the Science Museum! “You, meanwhile, will await our Foreign Office friend and explain to him that I have a lead that may indeed provide him a trustworthy answer. But I will have to strain my intellect to the utmost, if I am to bridge the gaps in our generation's understanding of the Analytical Engine, that it may become the means by which we reach our solution.“The name of this person who is constructing an Analytical Engine from Mr Babbage's notes? I shall enlist his aid.” “Hamilton Goodge is the name,” I told him. “A young man, well prepared in mathematics, and quite at home in his maze of levers and wheels. You know the type.” So I remained in the upper-story flat we shared at 221B Baker Street, and was able to give Mr Bond the news that Holmes's chase was afoot, this time in the workshop of a museum rather than in the moors of Scotland or the opium dens along the Thames. I went over my notes with him on the specifics Holmes had already noticed. And Holmes? Never a word from him the rest of the week. Bond went down on the Friday to see what was happening at the Museum, but Holmes refused to see him, so intent was he on formulating the exact instructions he must give the Analytical Engine in order to get the answers he needed. Which forms might tell us how the languages have changed? Choose at least one: husk flower leaf banana grass Some don't give evidence for anything. Some give evidence for what the ancestral language was like. Some give evidence for changes that have taken place.