As the first farmers spread agriculture across the globe, they also sowed seeds for most of today's languages, researchers from UCLA and the Australian National University conclude in the most up-to-date survey of research in the field.

"For the past two decades, scientists have debated whether farmers or hunter-gatherers are responsible for the spread of modern language families," said Jared Diamond, the study's lead author and a UCLA professor of geography and environmental health sciences. "It's been a real see-saw, but the preponderance of evidence now points to farmers with their encroaching culture as speakers of the ancestral languages from which most modern languages evolved."

"Because food production conferred enormous advantages on farmers compared to hunter-gatherers, it trigged outward dispersals of farming populations, bearing their languages and lifestyles," added co-author Peter Bellwood, an ANU professor of archaeology. "Over time, the farmers' languages tended to replace the languages of the hunter-gatherers whom they encountered."

The analysis, which appears in the April 25 issue of Science Magazine, shines new light on the origins of Japanese and several other Asian languages. But it does not resolve conflicts surrounding the Indo-European language family, including English, German, French, Italian and Spanish.

Diamond and Bellwood spent one year poring over 100 recent studies in archaeology, crop and livestock research, genetics, and linguistics. They compared the dates of the earliest-known agricultural remains with the dates when linguists believe specific languages diverged from each other. The UCLA-ANU team then looked at what DNA evidence reveals about the ancestry of today's speakers of languages that belong to 15 main ancestral language families.

Because these findings are so interdisciplinary, only a few scientists in the world possess the technical competence to make sense of them all. Diamond and Bellwood are rare in being equally fluent in linguistics, archaeology, plant and animal domestication, human skeletal remains, and modern human genetics.

Of the 15 language families studied, 12 language families - or 80 percent - appear to have spread at the same time as agricultural practices, the pair concludes.

The crucial role of farmers in spreading languages, the authors contend, helps explain why nearly one-half (seven of 15) of ancestral language families have evolved from the world's two major cradles of agriculture - China and the Fertile Crescent (today's Syria, Lebanon, Turkey, Jordan, Palestine, Iraq and Iran).

The link between the spread of language and agriculture also accounts for the tendencies of both to spread more rapidly in east-west directions, which have a greater likelihood of sharing the same temperatures and day-length plant germination schedules. Three of the most prominent language families, including the Indo-European family, have spread great distances - between 4,000 and 8,500 miles - along east-west directions in Europe, Asia, Africa and the Middle East, the authors point out.

Meanwhile, language families have rarely spread in a north-south direction, which would involve changes in day length and seasons, making crop spreads difficult, the authors said. In North and South America, for instance, no language family spans more than 2,000 miles in a north or south direction.

"Sites at the same latitude but different longitudes are likely to share not just climates but also agricultural systems," said Bellwood, author of "Prehistory of the Indo-Malaysian Archipelago" (University of Hawaii Press, 1997). "That means early farmers would have met fewer climatic obstacles as they spread in an east-west direction than in a north-south direction."

Still, farmers occasionally may have reverted to a hunter-gatherer lifestyle, especially if they migrated into regions that proved inhospitable for agricultural practices, the researchers contend. In other cases, intermarriage with hunter-gatherers appears to have diluted farmer genes. Either case would account for instances where hunter-gatherer DNA predominates among today's speakers of languages believed to have once spread through agriculture.

"There's a rough match between the languages and genes of these first farmers' descendents, but lots of people throw up their hands because of the exceptions," said Diamond, the Pulitzer Prize-winning author of "Guns, Germs and Steel: The Fates of Human Societies" (W.W. Norton, 1997) and winner of the National Medal of Science. "Nevertheless, we found these exceptions prove the rule."

Diamond and Bellwood contend that previous analysis may also have been flawed by an assumption that farmers' languages continued to be spoken in their homelands long after they had migrated elsewhere. In fact, the farmers' languages seem to have disappeared in several key cases after a new language came to dominate their homelands. With the farmer's native tongue surviving only in adopted lands, the direct evidence for the language's origins disappeared in the farmer's homeland.

This "replacement theory" would help explain the origins of Japanese, which have long stumped linguists because the language appears to have diverged long before archaeological remains show that agriculture was introduced to Japan by Korean rice farmers. Yet, few relationships can be found between today's Japanese language and today's Korean language.

Diamond and Bellwood argue that the now-extinct Korean predecessor to today's Japanese - Koguryo - was spoken in Korea until the Koguryo Kingdom was conquered in 700 by the rival Sillans, whose language, linguists agree, was the predecessor of today's Korean. By the time of the demise of Koguryo, Korean farmers would have been living in modern-day Japan for almost 1,100 years.

In keeping with the replacement theory, the ancestral languages responsible for today's Thai, Cambodian, Vietnamese, Burmese, Malay and Tagalog may all have roots in south China, Diamond and Bellwood assert. They contend that the Chinese homeland roots of these Austro-Asiatic and Austronesian languages have simply been buried by the subsequent rise of the Sino-Tibetan language family from which such modern languages as Cantonese and Mandarin evolved.

The major language family on which evidence remains the most inconclusive, the authors concede, is the Indo-European group. Some evidence points to the language family's birth near Turkey, where archaeological evidence has shown that wheat-growing, shepherding and other common agricultural practices originated. Other evidence points to an origin among horse-riding nomadic people in the steppes above Europe's Black Sea.

"I wake up in the morning and think farmers spread Indo-European languages," Diamond said, "and by 6 p.m. I've changed my mind."