By James L. Franklin
   About 20 years ago, I bought my first copy of Chapman's revision of Roget's Thesaurus for its size.  After flipping pages of the several versions of Roget's in a local bookshop, I  thought that Chapman's revision (the fourth edition, of 1977) had a larger vocabulary and would be a richer reference tool.  And generally I used it in that  limited fashion when writing, looking to jog my memory when I was  reaching for a word or to suggest a more descriptive word.

   Back on the copy desk after a long stint as a  reporter, I find I'm finally using the thesaurus primarily as a  headline writing tool.  I think of it as nuclear research, hunting for the atoms  and fragments that can fit in a one-column head.

   But every now and then I stumble back toward  using the thesaurus as Roget and Chapman intended.  When I can't solve my
headline problem with an acceptable shorter word, my eye wanders over the  larger entries. With the slot editor looking sourly in my direction,  thinking that if I have to look it up, he probably doesn't want it in a headline,  I begin to think the problem with the head isn't the one word but my  approach to the head. Because this thesaurus is not organized in  dictionary form as a sort of synonym finder, I'm really working with multiple  streams of word associations, and often they give me another way to think about the head. In brief, the solution isn't always finding a  synonym or near relation for the word that doesn't fit.

   Peter Roget's original contribution was in  categorizing English into tables to words, a systematic organization of words, concrete and abstract. It may not be your way of looking at the world  (and Chapman revised it considerably, to make it more  "developmental-existential" and less Platonic).  I have tended to use the book from  the back, finding pointers to possibly helpful word lists in the index, rather  than using the table of word lists directly.

   For example, when the front page head was "US  seeks to establish base in Afghanistan" and the runover or jump head was one  column, I was first looking for a larger, more evocative word than base.  The connections I found under base were not especially rich,  because, it turns out, neither Roget nor Chapman gave much consideration to base  as a military term.  I didn't find the suggestions stronghold or redoubt  or strong point from the tables suggested in the index under base (which did lead me to station or post in the first reference, to installation in a nearby entry, and, in another connection from the index, to
headquarters or command post, none of which had the force I was seeking).  But by using  the larger word lists, I found more fruitful connections in the major  tables for warfare, attack, defense, combatant, and arms, none of which would have been suggested by using the small end of the telescope and focusing  on my initial problem, finding a better word than base.

   As it turns out, my problem evaporated when I  paid attention to the fact that Afghanistan would never fit in the head order.  The result:
US seeks
Afghan
base for
troops

   But the process is what's important, looking  at the wider neighborhood of language and taking a step back when the first  choice won't work.  I like this thesaurus precisely because it offers a  bigger playground in which to reconsider my ideas.

   The fifth edition is better printed than the fourth, with a more spacious layout, and boasts 325,000 words and phrases. 

This is another of the three books I keep at my side each night.

James L. Franklin is a copy  editor on the night desk at The Boston Globe.  He has been a writer and  editor at the Globe since 1971 and has covered religion, the environment, and the suburbs.