Definitions: Love and Marriage

Linguist Geoff Nunberg had a great piece about how political factions use dictionary definitions in taking their stances toward gay marriage. It's a great analysis, and one that transcends the specific topic. I found of particular interest the following:

A couple of months ago, the editors of the Oxford English Dictionary made some long-overdue revisions in the definitions for a bunch of gender-related words. Before then, the dictionary's definition of girlfriend in the meaning of "sweetheart" read "a man's favorite female companion," which would have precluded lesbians from having girlfriends in the romantic sense. And the old definition of love read, "That feeling of attachment which is based upon difference of sex. . . and which is the normal basis of marriage." So both words were given new definitions that would cover their use to refer to same-sex relationships.[1]

This is hardly a matter of rampant political correctness, or of giving the words a new meaning. It isn't as if the English language has ever ruled out talking about lesbians having girlfriends, much less prevented Shakespeare from describing a romantic attachment between two men with the word love. It's just that when the definitions were written, those sorts of relationships were officially invisible.

Oh, what a great point. People use language to make themselves comfortable, and set definitions to try and form the world in the way they wished it was, and not in the way it actually is. To point to "traditional" definitions is really to point to traditional prejudices and recorded wishful thinking. What axe edges are the dictionary editors grinding? So what we get is a true example of "begging the question": people take a stance based on social norms, and then use definitions created by those norms as proof that the opinion is correct. Might as well say that infant mortality is normal and "the way things should be" because in the past there was lots of infant mortality.

The Evolution of Language

Two papers in the journal Nature this week examine the evolution of language. One measured the frequency at which verbs become more regular (use a simple -ed ending for past tense) over time at a rate inversely proportionate to the square root of their usage frequency. The site Science Codex has an article on the paper with interviews with some of the involved researchers.

The other paper looks at why some words use similar word forms across the entire Indo-European language family while others appear as unrelated forms. According to a press release from Nature:

Mark Pagel and colleagues used a statistical modelling technique to analyse four Indo-European languages: English, Spanish, Russian and Greek, and compared this to a database of 200 fundamental vocabulary meanings in 87 languages. They found that across all 200 meanings, commonly used words, such as numbers, evolve much more slowly, suggesting that the frequency with which specific words are used affects their rate of replacement over thousands of years.

Again, and with some appeal to common sense, the conclusion is that frequent use cements the form more thoroughly through the act of repetition.