Using search engines to check spelling: a caveat

[oryx_and_crake]

(inspired by this discussion)

Using a web search engine to determine the right spelling of some word is a good idea, if you use it wisely. Even the spelling that generates 40 000 hits may still be a wrong one. And here's why.

1. The bulk of web-pages returned by the search engine may be not in Russian at all, but in Byelorussian or Ukrainian, and you may not be able to tell the difference. (Try to yandex "карова".)

2. You'd be amazed but a lot of native speakers cannot spell correctly in their own language. That's why "база даных" will return about 40 000 hits which does not make this spelling legitimate, not a one little bit. Think native English speakers and words like "there/they're/their".

3. The misspelling may be deliberate, see here, for example: http://en.wikipedia.org/wiki/Udaff. Widespread Runet (Russian Internet) slang uses deliberate misspelling in a certain way (e.g. words like зайчик and котик become зайчег and котег, привет becomes превед etc.)

4. Typos happen. There are about 140 million people in Russia; if a few million people type the same relatively short and very commonly used word (like самый/самых that was discussed recently) there is a good probability that a few thousands of them will make the same typo. Think "teh" or "pwned".

Comments

[trepang.livejournal.com]

in russian, it's корова.

[oryx_and_crake]

That was exactly my point. Are you sure you have read my posting carefully?

[ex-ex-zhuzh.livejournal.com]

If you use Google, add one of these strings to your query:

site:gutenberg.org
site:bartleby.com
site:lib.ru intitle:классика

The last one works for Russian, the other two are for English, obviously.

According to my back-of-the-napkin calculations, this increases accuracy of the google-spell-checker about thousandfold (though still doesn't make it perfect).

[oryx_and_crake]

Yes, this will probably help. However there is a rub too: many Russian classic writers lived in the XIX century and even earlier, therefore their spelling might sometimes be a trifle obsolete. But all the same, probably it is indeed a thousand times better than "зайчеги".

[ashalynd.livejournal.com]

I am using a similar method to validate my Dutch texts and usually it is pretty easy to understand by context whether the spelling / words agreement is correct or not. Probably anybody with intermediate (or higher) language skills does the same and will, in most cases, distinguish a good example from a bad one.

For example, if you see "карова" and the rest of the words looks barely known to you, then you are probably reading Belorussian site :)

[upthera44.livejournal.com]

Thanks for the post. I've occasionally used this method and hadn't thought of some of these pitfalls.

[fiachasorcha.livejournal.com]

from all these people/comments I'm starting to think that I'm weird for actually using my dictionary to check spelling...or maybe I'm just bad at typing in russian and it's faster to use something in hard copy

[oryx_and_crake]

Dictionary is fine but it may not have some modern words/expressions, and it normally doesn't tell you anything about which word is good for which context, or how widely a certain word is used vs. some other word. So search engines have their uses, one just has to be careful and not rely on them too much.