Linguists have two rival views of human language: prescriptive and descriptive.
Descriptive says: Never mind what's "right" or "wrong", here is how people actually speak. Language is defined by people, so let's describe that.
Prescriptive says: Here is how people should speak. People who speak differently are simply wrong. They are damaging the language and should be corrected.
Prescriptive is fine for grammar school teachers, but in my opinion, if you want to communicate effectively in the real world of grownups, you should figure out how other people use words and use them the same way. If people are hopelessly confused about what a word means, then it's best to avoid it unless you are prepared to lead a campaign to educate the masses. (People who read my blog on whether iSCSI is a form of NAS (here and here) won't be surprised to hear that I am a descriptivist.)
I love the American Heritage Dictionary (see my book list) because it is descriptive and often backs me up when people try to correct my grammar. For instance, I use data as a singular word, but some presciptivists insist that it is the plural of the Latin word datum. American Heritage, on the other hand, reports that 77% of their usage panel now accept data as a singular word. Language changes. If it didn't, we'd all still be speaking Latin, Sanskrit, or Proto-Indo-European. (Proto-Indo-European is the ancestor to both Latin and Sanskrit, and the American Heritage has a dictionary of Proto-Indo-European word roots. My favorite is deru or drue, which is the root of words like tree, true, durable, druid, and even tryst - presumably because trysts happen under trees.)
Google is great for probing today's usage. In my recent blog on data deduplication, I had to decide whether to use dedup, de-dup, dedupe, or de-dupe. I checked with Google and got these results:
| dedup: | 35,400 matches |
| de-dup: | 75,200 matches |
| dedupe: | 116,000 matches |
| de-dupe: | 789,000 matches |
De-dupe is the winner by a landslide, so that's what I used. To get accurate results, you have to use the exact phrase feature of advance search, because otherwise Google treats the dash as a space.
You don't always have to go with the majority. Perhaps you want to invent a new phrase, or make a point. For instance, NetApp officially refers to iSCSI as IP SAN, even though Google shows iSCSI winning by 4,940,000 to 637,000. We want to make the point that people can use iSCSI in place of Fibre Channel SAN for surprisingly high-end applications. On the other hand, we always use iSCSI nearby so that people will know what we are talking about.
Sometimes you want a unique name. Last week my wife and I had a baby girl. Her name is Mira Hitz, which has no matches now but will as soon as I post this!


Comments