I don’t Care About Spellng

But Maybe You Should

Posted by Cerstin Mahlow on May 16, 2021

Do you check your writing before submitting a draft or hitting send for this very important mail? Do you trust your spell checker? Does it bother you when your word processor throws green and red squiggles over your text?

Do you care when an official message or notification you get contains spelling errors? Do spelling errors influence your judgement about the quality of a text?

There are various approaches or personal attitudes towards spelling. Spelling and grammar rules set prescriptive norms. If one refuses to adhere to norms in general, one might argue that content is king, and form and spelling are less important. And anyway, languages evolve and maybe what is judged wrong today will become the norm later. So why all this fuss about spelling correctly?

Creativity of course involves breaking rules, but most often this means deliberate play with rules and norms, so one has to know about those rules. And also the audience has to know about the rules to realize these creative ruptures. So yes, you should know about spelling and learn those rules.

But the computer takes care of it, doesn’t it? So you certainly don’t have to learn all these rules! Spell checkers claim to check and correct your spelling (and grammar), but as has been shown again and again, these checkers have limits and the most embarrassing errors will remain.

Receiving messages and texts full of errors leaves me with the impression that the sender/author didn’t care about language, didn’t care about the text, and also didn’t care about me, the recipient or audience. Yes, I do consider messages with spelling errors (that even a spell checker could have found!) impolite. There, I said it. And the worst texts are the ones that come with red and green squiggles still visible–some people not only don’t care about spelling, but also cannot handle their tools.

However, there are more reasons why we should create texts that use correct spelling and grammar: Automatic translation produces better results when the source text is error-free. Screen readers and other accessibility tools and generic browser extensions require correct texts; otherwise they fail or produce nonsense. NLP techniques use all kinds of texts as basis for the generation of language models; features of those texts including spelling and grammar will thus be propagated to future applications. What would you think about a chatbot whose answers contain spelling errors?